[HN Gopher] Predictions Scorecard, 2025 January 01
___________________________________________________________________
Predictions Scorecard, 2025 January 01
Author : timr
Score : 232 points
Date : 2025-01-10 00:27 UTC (22 hours ago)
(HTM) web link (rodneybrooks.com)
(TXT) w3m dump (rodneybrooks.com)
| rexreed wrote:
| I like Rodney Brooks, but I find the way he does these
| predictions to be very obtuse and subject to a lot of self-
| congratulatory interpretation. Highlighting something green that
| is "NET2021" and then saying he was right when something happened
| or didn't happen, when something related happened in 2024 mean
| that he predicted it right or wrong, or is everything subject to
| arbitrary interpretation? Where are the bold predictions? Sounds
| like a lot of fairly obvious predictions with a lot of wiggle
| room to determine if right or wrong.
| gcr wrote:
| NET2021 means that he predicted that the event would take place
| on or after 2021, so happening in 2024 satisfies that. Keep in
| mind these are six-year-old predictions.
|
| Are you wishing that he had tighter confidence intervals?
| rexreed wrote:
| If the predictions are meant to be bold, then yes. If they're
| meant to be fairly obvious, then no.
|
| For example, saying that flying cars will be in widespread
| use NET 2025 is not much of a prediction. I think we can all
| say that if flying cars will be in widespread use, it will
| happen No Earlier Than 2025. It could happen in 2060, and
| that NET 2025 prediction would still be true. He could mark
| it green in 2026 and say he was right, that, yes, there are
| no flying cars, and so mark his scorecard another point in
| the correct column. But is that really a prediction?
|
| A bolder prediction would be, say "Within 1-2 yrs of XX".
|
| So what is Rodney Brooks really trying to predict and say?
| I'd rather read about what the necessary gating conditions
| are for something significant and prediction-worthy to occur,
| or what the intractable problems are that would make
| something not be possible within a predicted time, rather
| than reading about him complain about how much overhype and
| media sensation there is in the AI and robotics (and space)
| fields. Yes, there is, but that's not much of a prediction or
| statement either, as it's fairly obvious.
|
| There's also a bit of an undercurrent of complaint in this
| long article about how the not-as-sexy or hyped work he has
| done for all those years has gone relatively unrewarded and
| "undeserving types" are getting all the attention (and
| money). And as such, many of the predictions and commentary
| on them read more as rant than as prediction.
| Denzel wrote:
| Presumably you read the section where Brooks highlights all
| the forecasts executives were making in 2017? His NET
| predictions act as a sort of counter-prediction to those
| types of blind optimistic, overly confident assertions.
|
| In that context, I'd say his predictions are neither
| obvious nor lacking boldness when we have influential
| people running around claiming that AGI is here today, AI
| agents will enter the workforce this year, and we should be
| prepared for AI-enabled layoffs.
| riffraff wrote:
| The NET estimation is supposed to be a counter to the
| irrational exuberance of media and PR. E.g. musk says
| they'll get humans to Mars in 2020, and the counter is "I
| don't think that will happen until at least 2030".
| kragen wrote:
| "NET2021" means "no earlier than 2021". So, if nothing even
| arguably similar happened until 2024, that sounds like a very
| correct prediction.
|
| Whether that's worth congratulating him about depends on how
| obvious it was, but I think you really need to measure "fairly
| obvious" at the time the prediction is made, not seven years
| later. A lot of things that seem "fairly obvious" now weren't
| obvious at all then.
| dang wrote:
| Related. Others?
|
| _Rodney Brooks Predictions Scorecard_ -
| https://news.ycombinator.com/item?id=34477124 - Jan 2023 (41
| comments)
|
| _Predictions Scorecard, 2021 January 01_ -
| https://news.ycombinator.com/item?id=25706436 - Jan 2021 (12
| comments)
|
| _Predictions Scorecard_ -
| https://news.ycombinator.com/item?id=18889719 - Jan 2019 (4
| comments)
| sashank_1509 wrote:
| Feels too self-congratulatory when he claims to be correct about
| self driving in the Waymo case. The bar he set is so broad and
| ambiguous, that probably anything Waymo did, would not qualify as
| self driving to him. So he think humans are intervening once
| every 1-2 miles to train the Waymo, we're not even sure if that
| is true, I heard from friends that it was 100+ miles but let us
| say Waymo comes out and says it is 1000 miles.
|
| Then I bet Rodney can just fiddle with goal post and say that
| 3.26 trillion miles were driven in US in 2024, and having a human
| intervene 1000 miles would mean 3.26 billion interventions, and
| that this is clearly not self driving. In fact until Waymo
| disables Internet on all cars and prices it never needs any
| intervention ever, Rodney can claim he's right, even then maybe
| not stopping exactly where Rodney wanted it to, might be proof
| that self driving doesn't work.
|
| Next big thing after deep learning prediction is clearly false.
| LLM is deep learning, scaled up, we are not in any sense looking
| past deep learning. Rodney I bet wanted it to be symbolic AI, but
| that is most likely a dead end, and the bitter lesson actually
| holds. In fact we have been riding this deep learning wave since
| Alex-Net 2012. OpenAI talked about scaling since 2016 and during
| that time the naysayers could be very confident and claim we
| needed something more, but OpenAI went ahead and proved out the
| scaling hypothesis and passed the language Turing test. We
| haven't needed anything more except scale and reasoning has also
| turned out to be similar. Just an LLM trained to reason, no
| symbolic merger, not even a search step it seems like.
| mvdtnz wrote:
| Your objection to him claiming a win on self driving is that
| you think that we can still define cars as self driving even
| when humans are operating them? Ok I disagree. If humans are
| operating them then they simply are not self driving by any
| sensible definition.
| sashank_1509 wrote:
| Human interventions are some non zero number in current self
| driving cars and will likely be that way for a while. Does
| this mean self driving is a scam and in fact it is just a
| human driving, and that these are actually ADAS. Maybe in
| some pedantic sense, you are right but then your definition
| is not useful, since it lumps cruise control/ lane-keeping
| ADAS and Waymo's in the same category. Waymo is genuinely,
| qualitatively a big improvement above any ADAS/ self driving
| system that we have seen. I suspect Rodney did not predict
| even Waymo's to be possible, but gave himself enough leeway
| so that he can pedantically argue that Waymo's are just ADAS
| and that his prediction was right.
| mvdtnz wrote:
| No one said scam (although in the case of Tesla it
| absolutely is). It's just not a solved problem yet.
| jdminhbg wrote:
| > It's just not a solved problem yet.
|
| Human driving isn't a solved problem either; the
| difference is that when a human driver needs intervention
| it just crashes.
| tsimionescu wrote:
| This is not about crashes. By all accounts, the Waymo
| cars are mostly fully self driving, I beleive even the
| article author agrees with that. This includes crash
| avoidance, to the extent that they can.
|
| The remote operation seems to be more about navigational
| issues and reading the road conditions. Things like
| accidentally looping, or not knowing how to proceed with
| an unexpected obstacle. Things that don't really happen
| to human drivers, even the greenest of new drivers.
| jdminhbg wrote:
| Ok, but crashes are much worse than navigational issues
| or accidentally looping. It's only status quo bias that
| makes us think driving is more solved if you get the
| accidental looping fixed before the crashing.
| mvdtnz wrote:
| Irrelevant.
| skywhopper wrote:
| Some of them are scams, yes. For stuff like Waymo, it
| definitely doesn't match the hype at the time he made the
| original predictions. As pointed out above, there were
| people in 2016 claiming we'd be buying cars without
| steering wheels that could go between any two points
| connected by roads by now.
| Spivak wrote:
| Yeah, I think semi-autonomous vehicles are a huge milestone
| and should be celebrated but the jump from semi-autonomous to
| fully-autonomous will, I think, feel noticeably different. It
| will be a moment future generations have trouble imagining a
| world where drunk or tired driving was ever even an issue.
| fragmede wrote:
| The future is here, just unevenly distributed. There are
| already people that don't have that issue, thanks to
| technology. That technology might be Waymo and not driving
| in the first place, or the technology might be smartphones
| and the Internet, which enables Uber/Lyft to operate. Some
| of them might use older technologies like concrete which
| enables people to live more densely and not have to drive
| to get to the nearest liquor establishment.
| munchler wrote:
| You can make exactly the opposite argument as well: You think
| that we can still define cars as human-driven even when they
| have self-driving features (e.g. lane keeping). If the car is
| self-driving in even the smallest way, then they simply are
| not human-operated by any sensible definition.
| skywhopper wrote:
| No one is making predictions or selling stock in the amount
| of "fully human controlled" vehicles.
| gwern wrote:
| The Waymo criticisms are absurd to the point of dishonesty. He
| criticizes a Waymo for... not pulling out fast enough around a
| truck, or for human criminals vandalizing them? Oh no, once
| some Waymos did a weird thing where they honked for a while!
| And a couple times they got stuck over a few million miles!
| This is an amazingly lame waste of space, and the fact that he
| does his best to only talk about Tesla instead of Waymo
| emphasizes how weak his arguments are, particularly in
| comparison to his earliest predictions. (Obviously only the
| best self-driving car matters to whether self-driving cars have
| been created.)
|
| "Nothing ever happens"... until it does, and it seems Brooks's
| prediction roundups can now be conveniently replaced with a
| little rock on it with "nothing in AI ever works" written on it
| without anything of value being lost.
| davedx wrote:
| Interesting, that wasn't my takeaway from the article at all!
|
| Direct quote from the article:
|
| > Then I will weave them together to explain how it is still
| pretty much business as usual, and I mean that in a good way,
| with steady progress on both the science and engineering of
| AI.
|
| There are some extremely emotional defences of Waymo on this
| comment thread. I don't quite understand why? Are they
| somehow inviolable to constructive criticism in the SV crowd?
| elicksaur wrote:
| It's interesting that in my reading of the post I felt like
| he hardly talked about Tesla at all.
|
| He calls out that Tesla FSD has been "next year" for 11
| years, but then the vast majority of the self-driving car
| section is about Cruise and Waymo. He also minorly mentions
| Tesla's promise of a robotaxi service and how it is unlikely
| to be materially different than Cruise/Waymo. The amount of
| space allocated to each made sense as I read it.
|
| For the meat of the issue: I can regularly drive places
| without someone else intervening. If someone else had to
| intervene in my driving 1/100 miles, even 1/1000 miles, most
| would probably say I shouldn't have a license.
|
| Yes, getting stuck behind a parked car or similar scenario is
| a critical flaw. It seems simple and non-important because it
| is not dangerous, but it means the drive would not be
| completed without a human. If I couldn't drive to work
| because there was a parked car on my home street, again,
| people would question whether I should be on the road, and
| I'd probably be fired.
| laweijfmvo wrote:
| Waymo is the best driver I've ridden with. Yes it has limited
| coverage. Maybe humans are intervening, but unless someone can
| prove that humans are intervening multiple times per ride,
| "self driving" is here, IMO, as of 2024.
| Denzel wrote:
| In what sense is self-driving "here" if the economics alone
| prove that it can't get "here"? It's not just limited
| coverage, it's practically non-existent coverage, both
| nationally and globally, with no evidence that the system can
| generalize, profitably, outside the limited areas it's
| currently in.
| AlotOfReading wrote:
| It's covering significant areas of 3 major metros, and the
| core of one minor, with testing deployments in several
| other major metros. Considering the top 10 metros are >70%
| of the US ridehail market, that seems like a long way
| beyond "non-existent" coverage nationally.
| ivanbalepin wrote:
| Speaking for one of those metro areas I'm familiar with:
| maybe in SF city limits specifically (where they still
| are half the Uber's share), but that's 10% of the
| population of the Bay Area metro. I'm very much looking
| forward to the day when I can take a robo cab from where
| I live near Google to the airport - preferably, much
| cheaper than today's absurd Uber rates - but today it's
| just not present in the lives of about 95+% of Bay Area
| residents.
| AlotOfReading wrote:
| Waymo's current operational area in the bay runs from
| Sunnyvale to fisherman's wharf. I don't know how many
| people that is, but I'm pretty comfortable calling it a
| big chunk of the bay.
|
| They don't run to SFO because SF hasn't approved them for
| airport service.
| kccqzy wrote:
| I just opened the Waymo app and its service certainly
| doesn't extend to Sunnyvale. I just recently had an
| experience where I got a Waymo to drive me to a Caltrain
| station so I can actually get to Sunnyvale.
| AlotOfReading wrote:
| The public area is SF to Daly City. The employee-only
| area runs down the rest of the peninsula. Both of them
| together are the operational area.
|
| Waymo's app only shows the areas accessible to you.
| Different users can have different accessible areas,
| though in the Bay area it's currently just the two
| divisions I'm aware of.
| riffraff wrote:
| Why would you consider the employee-only area? For that
| categorization to exist it must mean it's either
| unreliable for customers or too expensive cause there's
| too much human drivers on the loop. Either way it would
| not be considered as an area served by self driving, imo.
| AlotOfReading wrote:
| There are alternative possibilities, like "we don't have
| enough vehicles to serve this area appropriately" or "we
| don't have statistical power to ensure this area meets
| safety standards even though it looks fine", and "there
| are missing features (like freeways) that would make
| public service uncompetitive in this area" to simply "the
| CPUC hasn't approved a fare area expansion".
|
| It's an area they're operating legally, so it's part of
| their operational area. It's not part of their public
| service area, which I'd call that instead.
| modeless wrote:
| I wish! In Palo Alto the cars have been driving around
| for more than a decade and you still can't hail one.
| Lately I see them much less often than I used to,
| actually. I don't think occasional internal-only testing
| qualifies as "operational".
| stouset wrote:
| > preferably, much cheaper than today's absurd Uber rates
|
| I just want to highlight that the only mechanism by which
| this eventually produces cheaper rates is by removing
| having to pay a human driver.
|
| I'm not one to forestall technological progress, but
| there are a huge number of people already living on the
| margins who will lose one of their few remaining options
| for income as this expands. AI will inevitably create
| jobs, but it's hard to see how it will--in the short term
| at least--do anything to help the enormous numbers of
| people who are going to be put out of work.
|
| I'm not saying we should stop the inevitable forward
| march of technology. But at the same time it's hard for
| me to "very much look forward to" the flip side of being
| able to take robocabs everywhere.
| dullcrisp wrote:
| Do you ever drive yourself or would you feel guilty not
| paying a driver?
| AlotOfReading wrote:
| People living on the margins is fundamentally a social
| problem, and we all know how amenable those are to
| technical solutions.
|
| Let's say AV development stops tomorrow though. Is
| continuing to grind workers down under the boot of the
| gig economy really a preferred solution here or just a
| way to avoid the difficult political discussion we need
| to have either way?
| stouset wrote:
| I'm not sure how I could have been more clear that I'm
| not suggesting we stop development on robotaxis or
| anything related to AI.
|
| All I'm asking is that we take a moment to reflect on the
| people who _won 't_ be winners. Which is going to be a
| hell of a lot of people. And right now there is
| absolutely zero plan for what to do when these folks have
| one of the few remaining opportunities taken away from
| them.
|
| As awful as the gig economy has been it's better than the
| "no economy" we're about to drive them to.
| stavros wrote:
| This is orthogonal. You're living in a society with no
| social safety net, one which leaves people with minimal
| options, and you're arguing for keeping at least those
| minimal options. Yes, that's better than nothing, but
| there are much better solutions.
|
| The US is one of the richest countries in the world, with
| all that wealth going to a few people. "Give everyone
| else a few scraps too!" is better than having nothing,
| but redistributing the wealth is better.
| stouset wrote:
| I agree.
|
| But this is the society we live in now. We don't live in
| one where we take care of those whose jobs have been
| displaced.
|
| I wish we did. But we don't. So it's hard for me to feel
| quite as excited these days for the next thing that will
| make the world worse for so many people, even if it is a
| technological marvel.
|
| Just between trucking and rideshare drivers we're talking
| over 10 million people. Maybe this will be the straw that
| breaks the camel's back and finally gets us to take
| better care of our neighbors.
| stavros wrote:
| Yeah but it doesn't work to on the one hand campaign for
| not taking rideshare jobs away from people on an online
| forum, and on the other say "that's the society we live
| in now". If you're going to be defeatist, just accept
| those jobs might go away. If not, campaign for wealth
| redistribution and social safety nets.
| kiba wrote:
| Public transit would also remove lot of jobs and yet
| nobody suggesting we shouldn't build more public transit
| because it will remove jobs.
|
| This is just coming from using what we already know how
| to do better.
| stouset wrote:
| Public transit has a fundamentally local impact. It takes
| away some jobs but also provides a lot of jobs for a wide
| variety of skills and skill levels. It simultaneously
| provides an enormous number of benefits to nearby
| populations, including increased safety and reduced
| traffic.
|
| Self-driving cars will be disruptive globally. So far
| they primarily drive employment in a small set of the
| technology industry. Yes, there are manufacturing jobs
| involved but those are overwhelmingly going to be jobs
| that were already building human-operated vehicles. Self-
| driving cars will save many lives. But not as many as
| public transit does (proportionally per user) And it is
| blindingly obvious they will make traffic worse.
| rrr_oh_man wrote:
| > preferably, much cheaper than today's absurd Uber rates
|
| You haven't paid attention to how VC companies work.
| danenania wrote:
| Waymo has approval to operate in San Mateo County so it's
| likely coming pretty soon.
| Denzel wrote:
| You're narrowing the market for self-driving to the
| ridehail market in the top 10 US metros. That's kinda
| moving the goal posts, my friend, and completely ignoring
| the promises made by self-driving companies.
|
| The promise has been that self-driving would replace
| driving _in general_ because it'd be safer, more
| economical, etc. The promise has been that you'd be able
| to send your autonomous car from city to city without a
| driver present, possibly to pick up your child from
| school, and bring them back home.
|
| In that sense, yes, Waymo is nonexistent. As the article
| author points out, _lifetime miles_ for "self-driving"
| vehicles (70M) accounts for less than 1% of _daily_
| driving miles in the US (9B).
|
| Even if we suspend that perspective, and look at the
| ride-hailing market, in 2018 Uber/Lyft accounted for
| ~1-2% of miles driven in the top 10 US metros. [1] So,
| Waymo is a tiny part of a tiny market in a single nation
| in the world.
|
| Self-driving isn't "here" in any meaningful sense and it
| won't be in the near-term. If it were, we'd see Alphabet
| pouring much more of its war chest into Waymo to capture
| what stands to be a multi-trillion dollar market. But
| they're not, so clearly they see the same risks that
| Brooks is highlighting.
|
| [1]: https://drive.google.com/file/d/1FIUskVkj9lsAnWJQ6kL
| hAhNoVLj...
| AlotOfReading wrote:
| There are, optimistically, significantly less than 10k
| Waymos operating today. There are a bit less than 300M
| registered vehicles in the US. If the entire US
| automotive production were devoted solely to Waymos, it'd
| still take years to produce enough vehicles to drive any
| meaningful percentage of the daily road miles in the US.
|
| I think that's a bit of a silly standard to set for
| hopefully obvious reasons.
| HenryBemis wrote:
| > ..is a tiny part of a tiny market in a single nation in
| the world.
|
| Calculator was a small device that was made in one tiny
| market in one nation in the world. Now we all got a
| couple of hardware ones in our desk drawers, and a couple
| software ones on each smartphone.
|
| If a driving car can perform 'well' (Your Definition May
| Vary - YDMV) in NY/Chicago/etc. then it can perform
| equally 'well' in London, Paris, Berlin, Brussels, etc.
| It's just that EU has stricter rules/regulations while US
| is more relaxed (thus innovation happens 'there' and not
| 'here' in the EU).
|
| When 'you guys' (US) nail self-driving, it will only be a
| matter of time til we (EU) allow it to cross the pond. I
| see this as a hockey-stick graph. We are still on the
| eraser/blade phase.
| dingnuts wrote:
| if you had read the F-ing article, which you clearly did
| not, you would see that you are committing the sin of
| exponentiation: assuming that all tech advances
| exponentially because microprocessor development did (for
| awhile).
|
| Development of this technology appears to be logarithmic,
| not exponential.
| YetAnotherNick wrote:
| Here in the product/research sense, which is the hardest
| bar to cross. Making it cheaper takes time but generally we
| have reduced cost of everything by orders of magnitude when
| manufacturing ramps up, and I don't think self driving
| hardware(sensors etc) would be any different.
| Denzel wrote:
| It's not even here in the product/research sense. First,
| as the author points out, it's better characterized as
| operator-assisted semi-autonomous driving in limited
| locations. That's great but far from autonomous driving.
|
| Secondly, if we throw a dart on a map: 1) what are the
| chances Waymo can deploy there, 2) how much money would
| they have to invest to deploy, and 3) how long would it
| take?
|
| Waymo is nowhere near a turn-key system where they can
| setup in any city without investing in the infrastructure
| underlying Waymo's system. See [1] which details the
| amount of manual work and coordination with local
| officials that Waymo has to do _per_ city.
|
| And that's just to deploy an operator-assisted semi-
| autonomous vehicle in the US. EU, China, and India aren't
| even on the roadmap yet. These locations will take many
| more billions worth of investment.
|
| Not to mention Waymo hasn't even addressed long-haul
| trucking, an industry ripe for automation that makes
| cold, calculated, rational business decisions based on
| economics. Waymo had a brief foray in the industry and
| then gave up. Because they haven't solved autonomous
| driving yet and it's not even on the horizon.
|
| Whereas we can drop most humans in any of these locations
| and they'll mostly figure it out within the week.
|
| Far more than lowering the cost, there are fundamental
| technological problems that remain unsolved.
|
| [1]: https://waymo.com/blog/2020/09/the-waymo-driver-
| handbook-map...
| jsnell wrote:
| Where's the economic proof of impossibility? As far as I
| know Waymo has not published any official numbers, and any
| third party unit profitability analysis is going to be so
| sensitive to assumptions about e.g. exact depreciation
| schedules and utilization percentages that the error bars
| would inevitably be straddling both sides of the break-even
| line.
|
| > with no evidence that the system can generalize,
| profitably, outside the limited areas it's currently in
|
| That argument doesn't seem horribly compelling given the
| regular expansions to new areas.
| Denzel wrote:
| Analyzing Alphabet's capital allocation decisions gives
| you all the evidence necessary.
|
| It's safe to assume that a company's ownership takes the
| decisions that they believe will maximize the value of
| their company. Therefore, we can look at Alphabet's
| capital allocation decisions, with respect to Waymo, to
| see what they think about Waymo's opportunity.
|
| In the past five years, Alphabet has spent >$100B to
| buyback their stock; retained ~100B in cash. In 2024,
| they issued their first dividend to investors and
| authorized up to $70B more in stock buybacks.
|
| Over that same time period they've invested <$5B in
| Waymo, and committed to investing $5B more over the next
| few years (no timeline was given).
|
| This tells us that Alphabet believes their money is
| better spent buying back their stock, paying back their
| investors, or sitting in the bank, when compared to
| investing more in Waymo.
|
| Either they believe Waymo's opportunity is too small
| (unlikely) to warrant further investment, or when
| adjusted for the remaining risk/uncertainty (research,
| technology, product, market, execution, etc) they feel
| the venture needs to be de-risked further before
| investing more.
| VirusNewbie wrote:
| >believes their money is better spent buying back their
| stock,
|
| Alphabet has to buy back their stock because of the
| massive amount of stock comp they award.
| davedx wrote:
| > Alphabet has to buy back their stock because of the
| massive amount of stock comp they award.
|
| Wait, really? They're a publically traded company; don't
| they just need to _issue new stock_ (the opposite of
| buying it back) to employees, who can then choose to sell
| it in the public market?
| yakz wrote:
| It's much better comp if the value of the stock goes
| _up_.
| bronco21016 wrote:
| Isn't there a point of diminishing returns? Let's assume
| they hand over $70B to Waymo today. Can Waymo even
| allocate that?
|
| I view the bottlenecks as two things. Producing the
| vehicles and establishing new markets.
|
| My understanding of the process with the vehicles is they
| acquire them then begin a lengthy process of retrofitting
| them. It seems the only way to improve (read: speed up)
| this process is to have a tightly integrated
| manufacturing partner. Does $70B buy that? I'm not sure.
|
| Next, to establish new markets... you need to secure
| people and real estate. Money is essential but this isn't
| a problem you can simply wave money at. You need to get
| boots on the ground, scout out locations meeting
| requirements, and begin the fuzzy process of hiring.
|
| I think Alphabet will allocate money as the operation
| scales. If they can prove viability in a few more markets
| the levers to open faster production of vehicles will be
| pulled.
| Denzel wrote:
| Yes, correct, you're restating the "risk/uncertainty" in
| the form of various concrete hypotheses. :)
|
| Within the context of the original discussion around
| whether self-driving is here, today, or not, I think we
| can definitively see it's _not_ here.
| mortehu wrote:
| To be clear, buying back stock is one of the ways they
| can invest in Waymo (and other business units).
|
| Since Alphabet buybacks mostly just offset employee stock
| compensation, the main thing they are getting for this
| money is employees.
| sashank_1509 wrote:
| I would prefer if they just give employee bonuses rather
| than this indirect form of compensation
| sashank_1509 wrote:
| This is just a quirk of the modern stock market
| capitalist system. Yes, stock buybacks are more lucrative
| than almost anything other than a blitz-scaling B2B SAAS.
| But for good of society, I would prefer if Alphabet spent
| their money developing new technologies and not on stock
| buybacks / dividends. If they think every tech is a waste
| of money, then give it to charity, not stock buybacks.
| That said, Alohabet does develop new technologies
| regularly. Their track record before 2012 is stellar,
| their track record now is good (Alphafold, Waymo,
| Tensorflow, TPU etc), and it is nowhere close to being
| the worst offender of stock buybacks (I'm looking at you
| Apple), but we should move away from stock price over
| everything as a mentality and force companies to use
| their profits for the common good.
| davedx wrote:
| That's a very hand wavy argument. How about starting
| here:
|
| > Mario Herger: Waymo is using around four NVIDIA H100
| GPUSs at a unit price of $10,000 per vehicle to cover the
| necessary computing requirements. The five lidars, 29
| cameras, 4 radars - adds another $40,000 - $50,000. This
| would put the cost of a current Waymo robotaxi at around
| $150,000
|
| There are definitely some numbers out there that allow us
| to estimate within some standard deviations how
| unprofitable Waymo is
| jsnell wrote:
| (That quote doesn't seem credible. It seems quite
| unlikely that Waymo would use H100s -- for one, they
| operate cars that predate the H100 release. And H100s
| sure as hell don't cost just $10k either.)
|
| You're not even making a handwavy argument. Sure, it
| might sound like a lot of money, but in terms of unit
| profitability it could mean anything at all depending on
| the other parameters. What really matters is a) how long
| a period that investment is depreciated over; b) what
| utilization the car gets (ot alternatively, how much
| revenue it generates); c) how much lower the operating
| costs are due to not needing to pay a driver.
|
| Like, if the car is depreciated over 5 years, it's
| basically guaranteed to be unit profitable. While if it
| has to be depreciated over just a year, it probably
| isn't.
|
| Do you know what those numbers actually are? I don't.
| sashank_1509 wrote:
| I know for a fact Waymo uses TPU's not GPU, maybe it is
| equivalent to 4 H100's but TPU vs GPU is somewhat apples
| vs oranges
| shrubble wrote:
| Does Wayne operate in heavy rain and any kind of snow or ice
| conditions?
| bhelkey wrote:
| The author specifically calls out that the taxi service
| needs not operate in all weather conditions or times of
| day.
|
| > First driverless "taxi" service in a major US city, with
| dedicated pick up and drop off points, and restrictions on
| weather and time of day.
|
| However, their analysis _this year_ is that, "This is
| unlikely to happen in the first half of this century."
|
| The prediction is clear. The evaluation is dishonest.
| benreesman wrote:
| Waymo cars can drive. Everything from the (limited) public
| literature to riding them personally has me totally persuaded
| that they can drive.
|
| DeepMind RL/MCTS can succeed in fairly open-ended settings like
| StarCraft and shit.
|
| Brain/DeepMind still knocks hard. They under-invested in LLMs
| and remain kind of half-hearted around it because they think
| it's a dumbass sideshow because it is a dumbass sideshow.
|
| They train on TPU which costs less than chips made of Rhodium
| like a rapper's sunglasses, they fixed the structural limits in
| TF2 and PyTorch via the Jax ecosystem.
|
| If I ever get interested in making some money again Google is
| the only FAANG outfit I'd look at.
| tylerflick wrote:
| I can tell you as someone that crosses paths almost everyday
| with a Waymo car, they absolutely due work. I would describe
| their driving behavior as very safe and overly cautious. I'm
| far more concerned of humans behind the wheel.
| benreesman wrote:
| I especially love how they can go _fast_ when it's safe and
| _slow_ when the error bars go up even a little.
|
| It's like being in the back seat of Nikki Lauda's car.
| Bootvis wrote:
| As shown here:
|
| https://www.youtube.com/watch?v=hVZ8NyV4pXU
| benreesman wrote:
| Perfect clip out of all of YouTube.
| fouronnes3 wrote:
| Does Waymo run on JAX?
| KKKKkkkk1 wrote:
| Which structural limits of TF2 and PyTorch were fixed via the
| Jax ecosystem?
| vessenes wrote:
| Agreed Waymo cars can drive. Also I don't believe that, say,
| when a city bus stops on a narrow street near a school
| crosswalk, that the decision to edge out and around it is
| made on board the car, as I saw recently. The "car" made the
| right decision, drove it perfectly, and was safe at all
| times, but I just don't think anyone but a human in a call
| center said yes to that.
| khafra wrote:
| > So he think humans are intervening once every 1-2 miles to
| train the Waymo
|
| Just to make sure we're applying our rubric fairly and
| universally: Has anyone else been in an Uber where you wished
| you were able to intervene in the driving a few times, or at
| least apply RLHF to the driver?
|
| (In other words: Waymo may be imperfect to the point where
| corrections are sometimes warranted; that doesn't mean they're
| not already driving at a superhuman level, for most humans.
| Just because there is no way for remote advisors to provide
| better decisions for human drivers doesn't mean that human-
| driven cars would not benefit from that, if it were
| available.).
| tsimionescu wrote:
| To apply this benchmark, you'd have to believe that Waymo is
| paying operators to improve the quality of the ride, not to
| make the ride possible at all. That is, you'd have to believe
| that the fully autonomous car works and gets you to your
| destination safely and in a timely manner (at the level of a
| median professional human driver), but Waymo decided that's
| not good enough and hired operators to improve beyond that.
| This seems very unlikely to me, and some of the (few)
| examples I've seen online were about correcting significant
| failures, such as waiting behind a parked truck indefinitely
| (as if it were stopped at a red light) or looping around
| aimlessly in a parking lot.
|
| You'd also have to believe that when you wished to change how
| your Uber driver drove, you'd actually have improved things
| rather than worsened them.
| lukeschlather wrote:
| Let's suppose Waymo's fully automated stuff has tenfold-
| fewer fatal collisions than a human. There's no way to
| avoid the fatal accidents a human causes, and the solution
| to Waymos getting stuck sometimes is simple. The point is
| that the Waymo can actually be described as superior to a
| human driver, and the fact that its errors can be corrected
| with review is a feature and not a bug - they optimize for
| those kinds of errors rather than unrecoverable ones.
| tsimionescu wrote:
| I think that, _if_ it were true that Waymo cars require human
| intervention every 1-2 miles (thus requiring 1 operator for
| every, say, 1-2 cars, probably constantly paying attention
| while the car is in motion), then it would be fair to say that
| the cars are not really self driving.
|
| However, if the real number is something like an intervention
| every 20 or 100 miles, and so an operator is likely passively
| monitoring dozens of cars, and the cars themselves ask for
| operator assistance rather than the operator actively
| monitoring them, then I would agree with you that Waymo has
| really achieved full self driving and his predicitons on the
| basic viability have turned out wrong.
|
| I have no idea though which is the case. I would be very
| interested if there are any reliable resources pointing one way
| or the other.
| skywhopper wrote:
| I disagree that regular interventions every two trips where
| you have no control over pickup or dropoff points counts as
| full self driving.
|
| But that definition doesn't even matter. The key factor is
| whether the additional overhead, whatever percentage it is,
| makes economic sense for the operator or the customer. And it
| seems pretty clear the economics aren't there yet.
| 4ndrewl wrote:
| Nonsense. If you spoke about self-driving cars a few decades
| ago you would have understood it to have meant that you could
| go to a dealer and buy a car that would drive itself, wherever
| you might be, without your input as a driver.
|
| No-one would have equated the phrase "we'll have self-driving
| cars" with "some taxis in a few of US cities"
| Schiendelman wrote:
| That's how all innovation works. Ford never said people asked
| for a faster horse, but the theory holds. It doesn't matter
| what benchmarks you set, the market finds an interesting way
| to satisfy people's needs.
| bhelkey wrote:
| The prediction is:
|
| > First driverless "taxi" service in a major US city, with
| dedicated pick up and drop off points, and restrictions on
| weather and time of day.
|
| Their 2025 analysis is: "This is unlikely to happen in the
| first half of this century."
|
| The prediction is clear. The evaluation is dishonest.
| throw-qqqqq wrote:
| I agree.. Waymo sells +150k rides every week according to
| Alphabet's Q3 2024 earnings announcement. Yes they need
| human assistance once in a while. I know of plenty other
| automation that needs to be tickled or rebooted
| periodically to work, that most would still say works
| automatically.
|
| Maybe he has a very narrow or strict definition of
| 'driverless'. That would explain the "not in this half of
| the century"-sentiment. I mean, it's 25 years!
| littlestymaar wrote:
| > when he claims to be correct about self driving in the Waymo
| case. The bar he set is so broad and ambiguous, that probably
| anything Waymo did, would not qualify as self driving to him
|
| Honestly, back in 2012 or something I was convinced that we
| would have autonomous driving by now, and by autonomous driving
| I definitely didn't mean "one company is able to offer
| autonomous taxi rides is a very limited amount of places with
| remote operator supervision", the marketing pitch has always
| been something along "the car you'll buy will be autonomously
| driving you to whatever destination you ask for, and you'll be
| just a passenger in you own car", and we definitely aren't
| there at all when all we have is Waymo.
| UniverseHacker wrote:
| It seems to me that the redefined flying cars for extremely
| wealthy people did happen? eVTOLs are being sold/delivered to the
| general public. Certainly still pretty rare, as I've never seen
| one in real life. I'd love to have one but would probably hate a
| world where everyone has them.
|
| Not really wanting to have this argument a second time in a week
| (seriously- just look at my past comments instead of replying
| here as I said all I care to say
| https://news.ycombinator.com/item?id=42588699), but he is totally
| wrong about LLMs just looking up answers in their weights- they
| can correctly answer questions about totally fabricated new
| scenarios, such as solving simple physics questions that require
| tracking the location of objects and reporting where they will
| likely end up based on modeling the interactions involved. If you
| absolutely must reply that I am wrong at least try it yourself
| first in a recent model like GPT-4o and post the prompt you
| tried.
| laweijfmvo wrote:
| Kobe Bryant basically commuted by helicopter, when it was
| convenient. It may have even taken off and landed at his house,
| but probably not exactly at all of his destinations. Is a
| "flying car" fundamentally that much different?
| UniverseHacker wrote:
| I think the difference is that a helicopter is extremely
| technical to fly requiring complex and expensive training,
| and the eVTOL is supposed to be extremely simple to fly. Also
| the eVTOL in principle is really cheap to make if you just
| consider the materials and construction costs- probably
| eventually much cheaper than a car.
|
| I was curious so I looked up how much you can buy the
| cheapest new helicopters for, and they are cheaper than an
| eVTOL right now- the XE composite is $68k new, and things
| like that can be ~25k used. I'm shocked one can in principle
| own a working helicopter for less than the price of a 7 year
| old Toyota Camry.
| xarope wrote:
| But I'm sure running costs (aviation fuel), hanger costs,
| maintenance costs, cost to maintain pilot license are far
| more expensive, compared to driving a car.
| torginus wrote:
| Nothing that flies in the air is that safe for its
| passengers or its surroundings - not without restrictions
| placed on it and having a maintenance schedule that most
| people would not be comfortable following.
|
| Most components are safety critical in ways that their
| failure can lead to an outright crash or feeding the pilot
| false information leading him to make a fatal mistake. Most
| cars can be run relatively safely even with major
| mechanical issues, but something as 'simple' as a broken
| heater on a pitot tube (or any other component) can lead to
| a crash.
|
| Then there's an issue of weather - altitude, temperature,
| humidity, wind speed can create an environment that makes
| it either impossible, unsafe, or extremely unpleasant -
| imagine flying into an eddy current that stalls out the
| aircraft, making your ass drop a few feet.
|
| Flying's a nice hobby, and I have great respect to people
| who can make a career out of it, but I'd definitely not get
| into these auto-piloted eVTOLs, nor should people who don't
| know what they are doing.
|
| Edit: Also unlike helicopters, which can autorotate, and
| fixed wing aircraft, that can glide, eVTOLs just drop out
| of the sky.
| UniverseHacker wrote:
| I would expect eVTOLs to be capable of greater redundancy
| than a helicopter or fixed wing aircraft - with no single
| point of failure that could make them drop from the sky.
| It would add little weight to have two or more
| independent electrical and motor systems each capable of
| making a semi controlled landing on its own, but must
| coordinate to provide the full rated lift. Marketing
| materials claim the Blackfly has triple redundancy. I
| suppose one could have software logic glitches that cause
| all modular systems to respond inappropriately to
| conditions in unison.
| mgfist wrote:
| eVTOLs are going to be much more expensive to build than
| helicopters because they have far more stringent
| weight/strength requirements due to low battery energy
| density (relative to aviation fuel).
|
| The idea is to have far cheaper operating costs. Electric
| motors are far more efficient than ICE, so you should have
| much cheaper energy costs. Electric motors are also simpler
| than ICE so you should have cheaper maintenance with less
| required downtime compared to helicopters.
|
| Of course, most of this is still being tested and worked
| on. But we are getting closer to having these get certified
| (FAA just released the SFAR for eVTOL, the first one since
| the 1940s).
| xarope wrote:
| I had a friend who used to (still does) fly RC helicopters;
| that requires quite a bit of skill. Meanwhile, I think
| anybody can fly a DJI drone. I think that's what will
| transform "flying" when anybody, not just a highly skilled
| pilot, can "drive" a flying car (assuming it can be as safe
| as a normal car... which somehow I doubt)
| input_sh wrote:
| Can you imagine thousands of flying cars flying low over
| urban areas?
|
| Skill level needed for "driving" would increase by a lot,
| noise levels would be abysmal, security implications would be
| severe (be they intentional or mechanical in nature), privacy
| implications would result in nobody wanting to have windows.
|
| This is all more-or-less true for drones as well, but their
| weight is comparable to a todler, not to a polar bear. I
| firmly believe they'll never reach mass usage, but not
| because they're impossible to make.
| UniverseHacker wrote:
| That does sound truly awful. I already hate the noise of
| internal combustion cars and am looking forward to cars
| getting quieter.
| Al-Khwarizmi wrote:
| Yeah, as an NLP researcher I was reading the post with interest
| until I found that gross oversimplification about LLMs, which
| has been repeatedly proved wrong. Now I don't trust the
| comments and predictions on the other fields I know much less
| about.
| Animats wrote:
| > That being said, we are not on the verge of replacing and
| eliminating humans in either white collar jobs or blue collar
| jobs.
|
| Tell that to someone laid off when replaced by some "AI" system.
|
| > Waymo not autonomous enough
|
| It's not clear how often Waymo cars need remote attention, but
| it's not every 1-2 miles. Customers would notice the vehicle
| being stopped and stuck during the wait for customer service.
| There are many videos of people driving in Waymos for hours
| without any sign of a situation that required remote
| intervention.
|
| Tesla and Baidu do use remote drivers.
|
| The situations where Waymo cars get stuck are now somewhat
| obscure cases. Yesterday, the new mayor of SF had two limos
| double-parked, and a Waymo got stuck behind that. A Waymo got
| stuck in a parade that hadn't been listed on Muni's street
| closure list.
|
| > Flying cars
|
| Probably at the 2028 Olympics in Los Angeles. They won't be cost-
| effective, but it will be a cool demo. EHang recently put solid
| state batteries into their flying car and got 48 minutes of
| flight time, instead of their previous 25 minutes. Ehang is
| basically a scaled-up quadrotor drone, with 16 motors and props.
| EHang been flying for years, but not for very long per recharge.
| Better batteries will help a lot.
|
| [1] https://aerospaceamerica.aiaa.org/electric-air-taxi-
| flights-...
| shlomo_z wrote:
| > Tell that to someone laid off when replaced by some "AI"
| system. What are some good examples? I am very skeptical of
| anyone losing their jobs to AI. People are getting laid off for
| various reasons: - Companies are replacing American tech jobs
| with foreigners - Many companies hired more devs than they need
| - companies hired many devs during the pandemic, and don't need
| them anymore
|
| Some companies may claim they are replacing devs with AI. I
| take it with a grain of salt. I believe some devs were probably
| replaced by AI, but not a large amount.
|
| I think there may be a lot more layoffs in the future, but AI
| will probably account for a very small fraction of those.
| theLiminator wrote:
| I imagine there aren't really layoffs, but slowing/stopping
| of hiring as you get more productivity out of existing devs.
| I imagine in the future, lots of companies will just let
| their employee base slowly attrition away.
| lolinder wrote:
| > I believe some devs were probably replaced by AI, but not a
| large amount.
|
| I'm not even sold on the idea that there were _any_. The
| media likes to blame AI for the developer layoffs because it
| makes a much more exciting story than interest rates and
| arcane tax code changes.
|
| But the fact is that we don't need more than the Section 174
| changes and the end of ZIRP to explain what's happened in
| tech. Federal economic policy was set up to direct massive
| amounts of investment into software development. Now it's
| not. That's a _real_ , quantifiable impact that can readily
| explain what we've seen in a way that the current
| productivity gains from these tools simply can't.
|
| Now, I'll definitely accept that many companies are
| _attributing_ their layoffs to AI, but that 's for much the
| same reason that the media laps the story up: it's a far
| better line to feed investors than that the financial
| environment has changed for the worse.
| Mistletoe wrote:
| But I see so many devs typing here saying how vital AI is
| to their writing code efficiently and quickly now. If that
| is true then you need way less devs. Are people just
| sitting idle at their desks? I do see quite a bit of tech
| layoffs for sure. Are you saying devs aren't part of the
| workers being laid off?
|
| >In 2024: At least 95,667 workers at U.S.-based tech
| companies have lost their jobs so far in the year,
| according to a Crunchbase News tally.
| baq wrote:
| Induced demand means we'll need more devs than we have
| right now since every dev can produce more value (anyone
| using cursor for a longer while should be able to confirm
| that easily).
|
| The problem is different in the meantime: nobody wants to
| be paying for training of those new devs. Juniors don't
| have the experience to call LLM's bullshit and seniors
| don't get paid to teach them since LLMs replaced interns
| churning out boilerplate.
| tsimionescu wrote:
| > Are you saying devs aren't part of the workers being
| laid off?
|
| No, they are saying that the reason for the layoffs is
| not AI, it is financial changes making devs too
| expensive.
|
| > If that is true then you need way less devs.
|
| This does not follow. First of all, companies take a long
| time to measure dev output, it's not like you can look at
| a burn down chart over two sprints and decide to fire
| half the team because it seems they're working twice as
| fast. So any productivity gains will show up as layoffs
| only after a long time.
|
| Secondly, dev productivity is very rarely significantly
| bounded by how long boilerplate takes to write. Having a
| more efficient way to write boilerplate, even massively
| more efficient, say 8h down to 1h, will only marginally
| improve your overall throughput, at least at the senior
| level: all that does is free you to think more about the
| complex issues you needed to solve. So if the task would
| have previously taken you 10 days, of which one day was
| spent on boilerplate, it may now take you, say, 8-9 days,
| because you've saved one day on boilerplate, plus some
| more minor gains here and there. So far from firing 7 out
| of every 8 devs, the 8h-to-1h boilerplate solution might
| allow you to fire 1 dev in a team of 10.
| jdminhbg wrote:
| > But I see so many devs typing here saying how vital AI
| is to their writing code efficiently and quickly now. If
| that is true then you need way less devs.
|
| Sure, in the same sense that editors and compilers mean
| you need way less devs.
| guappa wrote:
| A lot of devs are hacks. If an AI can do your job you had
| no value as a software developer.
| lukeschlather wrote:
| https://www.bls.gov/ooh/computer-and-information-
| technology/...
|
| BLS reports ~1.9 million software developer jobs and
| predicts 17% growth through 2033. Crunchbase is talking
| about "tech workers" not developers. And they don't even
| say that tech employment is down. I predict that when BLS
| publishes their preliminary job numbers for 2024 it will
| be at least 1.85 million, not 1.9 million as suggested by
| your Crunchbase News. I would lay 2:1 odds that it will
| be higher than 2023's number.
| skywhopper wrote:
| It just isn't true that AI has made developers more
| efficient. Some might claim such on this site, but the
| vast majority of developers aren't using it, or they find
| it to be a drag on their productivity (because for most
| tasks the median software engineer has to do, it actually
| can't help), and the ones that do use it are (unknowingly
| maybe) exaggerating its impact.
|
| Devs are getting laid off, yes. AI is not the reason.
| Executive/shareholder priorities are the reason.
| thefaux wrote:
| Why would Jevon's paradox not apply to human labor?
|
| I am not sure what I expect for software developers
| besides that the nature if the work will change but it is
| still too early to say exactly how. We certainly cannot
| extrapolate linearly or exponentially from the past few
| years.
| lolinder wrote:
| > Are you saying devs aren't part of the workers being
| laid off?
|
| Of course not. The Section 174 changes are really only
| relevant to software devs--the conversation in the months
| leading up to them kicking in was all about how it would
| kill software jobs. But then when it happened the media
| latched onto this idea that it was the result of
| automation, with zero evidence besides the timing.
|
| Since the timing also coincided with a gigantically
| important change to the tax code and a rapid increase in
| interest rates, both of which were predicted to kill
| software jobs, I'm suggesting that blaming AI is silly--
| we have a proximate cause already that is much more
| probable.
| littlestymaar wrote:
| > But I see so many devs typing here saying how vital AI
| is to their writing code efficiently and quickly now. If
| that is true then you need way less devs
|
| Same can be said for github, and open-source deoendency
| management tools like npm, and I'd argue that it had an
| even a much bigger impact then, and did you see what
| happen afterwards? Where were the mass layoffs back then?
| The number of software developers is actually much higher
| than before that era.
| rcpt wrote:
| It's pretty much impossible to get work as a copywriter now
| RamblingCTO wrote:
| I was thinking about this. I think we have an
| overcorrection right now. People get laid off because of
| expected performance of AI, not real performance. With
| copywriting and software development we have three options:
|
| 1. leaders notice they were wrong, start to increase human
| headcount again 2. human work is seen as boutique and
| premium, used for marketing and market placement 3. we just
| accept the sub-par quality of AI and go with it (quite
| likely with copywriting I guess)
|
| I'd like to compare it with cinema and Netflix. There was a
| time where lost of stuff was mindless shit, but there was
| still place for A24 and it took the world by storm. What's
| gonna happen? No one knows.
|
| But anyway, I figure that 90% of "laid off because of AI"
| is just regular lay-offs with a nice sounding reason. You
| don't loose anything by saying that and only gain in
| stakeholder trust.
| harvodex wrote:
| 90% might even be too low.
|
| If you look up business analyst type jobs on JP Morgan
| website they are still hiring a ton right now.
|
| What you actually notice is how many are being outsourced
| to other countries outside the US.
|
| I think the main process at work is 1% actual AI
| automation and a huge amount of return to the office in
| the US while offshoring the remote work under the cover
| of "AI".
| davedx wrote:
| Yeah, the AgentForce thing is a classic example. Internal
| leaks say Salesforce is using it as cover for more regular
| (cost cutting based) layoffs. People who've actually
| evaluated AgentForce don't think it's ready for prime time.
| It's more smoke and mirrors (and lots of marketing).
| brcmthrowaway wrote:
| What is the silver bullet for battery tech?
| Animats wrote:
| Solid state batteries. Prototypes work, but high-volume
| manufacturing doesn't work yet. The major battery
| manufacturers are all trying to get this to production. Early
| versions will probably be expensive.
|
| Maybe a 2x improvement in kwh/kg. Much less risk of fire or
| thermal runaway. Charging in < 10 mins.
| Teever wrote:
| The one thing I'm curious about with solid state batteries
| is if there's a path towards incremental improvements in
| power density like we've seen with lithium batteries?
|
| It would be unfortunate if we get solid state batteries
| that have the great features that you describe but they're
| limited to 2x or so power density. Twice the power density
| opens a lot of doors for technology improvements and
| innovation but it's still limiting for really cool things
| like humanoid robotics and large scale battery powered
| aircraft.
| Animats wrote:
| Somebody may come up with a new battery chemistry. There
| are many people trying. There are constraints other than
| energy density - charge rate, discharge rate, safety,
| lifetime, cooling, etc. Lithium-air batteries have an
| energy density which potentially approaches that of
| gasoline, but decades of work have not produced anything
| usable.[1]
|
| There are, of course, small startups promising usable
| lithium-air batteries Real Soon Now.[2]
|
| [1]
| https://en.wikipedia.org/wiki/Lithium%E2%80%93air_battery
|
| [2] https://airenergyllc.com/
| mgfist wrote:
| For aircraft, hydrogen is the better option. Batteries
| are fine for very short hauls (100 miles). Beyond that
| hydrogen is quite good.
| brcmthrowaway wrote:
| What about LK-99? Twitter influencers were talking about
| that.
| adgjlsfhk1 wrote:
| I think there are ~3 major battery improvements to watch out
| for.
|
| 1. Solid state batteries. Likely to be expensive, but promise
| better energy density.
|
| 2. Some really good grid storage battery. Likely made with
| iron or molten salt or something like that. Dirt cheap, but
| horrible energy density.
|
| 3. Continued Lithium ion battery improvements, e.g. cheaper,
| more durable etc.
| Animats wrote:
| There are now a few large flow batteries. Here's one that's
| 400 megawatt-hours.[1] Round trip efficiency is poor and
| the installation is bulky, but storage is just tanks of
| liquid that are constantly recycled.
|
| [1] https://newatlas.com/energy/worlds-largest-flow-
| battery-grid...
| davedx wrote:
| My money is on saltwater batteries. You can make them
| really cheaply. Flow batteries are still too complicated
| IMO.
| davedx wrote:
| I think what Waymo's achieved is really impressive, and I like
| the way they've rolled out (carefully), but there's a lot of
| non evidence based defense of them in this comment thread.
| YouTube videos of people driving for hours are textbook
| survivorship bias. (What about all the videos people made but
| didn't upload because their drive didn't go perfectly?)
|
| Nobody knows how many times operators intervene, because Waymo
| hasn't said. It's literally _impossible_ to deduce.
|
| Which means I also agree his estimate could also be wildly
| wrong too.
| skywhopper wrote:
| He's saying AI can't do the work of humans, not that dumb
| executives won't pretend it can.
| vikrantrathore wrote:
| For me this predictions are kind of being aware of how progress
| can happen based on history, but this will not lead to any
| breakthrough. I am not in the camp of being skeptic so I still
| like the hype cycle, they create an environment for people to
| break the boundaries and sometimes help untested ideas and things
| to be explored. This might not have happen if there is no hype
| cycle. I am in the camp of people who are positive as George
| Bernard Shaw in his 2 quotes: 1. A life spent
| making mistakes is not only more honorable, but more useful than
| a life spent doing nothing. 2. The reasonable person adapts
| themselves to the world: the unreasonable one persists in trying
| to adapt the world to themself. Therefore all progress depends on
| the unreasonable person. (Changed man to person as I feel it
| should be gender neutral)
|
| In hindsight when we look back, everything looks like we
| anticipated, so predictions are no different some pans out some
| doesn't. My feeling after reading prediction scorecard is that
| you need a right balance between risk averse (who are either
| doubtful or do not have faith things will happen quickly enough)
| and risk takers (one who is extremely positive) for anything good
| to happen. Both help humanity to move forward and are necessary
| part of nature.
|
| It is possible AGI might replace humans in a short term and then
| new kind of work emerges and humans again find something
| different. There is always a disruption with new changes and some
| survive and some can't, even if nothing much happens its worth
| trying as said in quote 1.
| sgt101 wrote:
| I feel a counter is that hyping and going along with hype leads
| to substantial misallocation of capital and this leads to human
| misery.
|
| How much money has been burned on robo-taxis which could have
| been spent on incubators for kids.
| zzzeek wrote:
| what a weird writer, lots of interesting things to talk about but
| this very long essay continued to circle back to being author-
| self-obsessed with their own prowess and drawing out huge
| expositions and bullet lists on how well they are at predicting
| things. Call it self-referential-appeal-to-authority.
| mberning wrote:
| It is odd. The product of a mind which clearly thinks very
| highly of itself.
| fifticon wrote:
| Another perspective is that it is a person who takes great
| care/is very thorough, to examine and re-evaluate his
| reasonings, and makes an effort to explain the logic in his
| reasoning, which can be helpful if you are trying to figure out
| if you agree or disagree.
| ynniv wrote:
| It's long, so I'm skimming a little and... flying cars. If you
| don't know why we don't have flying cars, you're not a good
| engineer.
|
| It really doesn't matter what prestigious lab you ran, as that
| apparently didn't impart the ability to think critically about
| engineering problems.
|
| [Hint: Flying takes 10x the energy of driving, and the
| cost/weight/volume of 1 MJ hasn't changed in close to a hundred
| years. Flying cars require a 10x energy breakthrough.]
| anon7000 wrote:
| The article is not optimistic on flying cars. The prediction is
| that an expensive flying car could be purchased no earlier than
| 2036, and notes a strong possibility that it won't even happen
| by 2050. Plus states that minor success (aka 0.1% of car sales
| are flying cars) isn't going to happen in his lifetime.
|
| The author also expands on this:
|
| > Don't hold your breath. They are not here. They are not
| coming soon.
|
| > Nothing has changed. Billions of dollars have been spent on
| this fantasy of personal flying cars. It is just that, a
| fantasy, largely fueled by spending by billionaires.
|
| It's worth actually reading the article before trashing
| someone's career and engineering skills!
| ynniv wrote:
| Engineering is about focusing on what matters. There's no
| point in talking about flying cars: they will exist when
| portable fusion exists, so just talk about that.
| cudgy wrote:
| So you are saying that a true engineer doesn't read
| articles and criticizes a successful engineer that wrote
| said article with hand-wavy arguments?
| tsimionescu wrote:
| The article is responding to claims by CEOs of car companies,
| industry and business press, and other hype sources that keep
| predicting flying cars next year or so. It's predicting that,
| against this hype, it will not come to pass. Not sure why
| you've worded your comment in such a way as if the article was
| hyping up flying cars.
|
| Not to mention, since we do have helicopters, the engineering
| challange of flying cars is almost entirely unrelated to energy
| costs (at least for the super rich, the equivalent of, say, a
| Rolls Royce, not of a Toyota). The thing stopping flying cars
| from existing is that it is extremely hard to make an easy to
| pilot flying vehicle, given the numerous degrees of freedom
| (and potential catastrophic failure modes); and the
| significantly higher impredictability and variance of the
| medium (air vs road surface).
|
| Plus, the major problem of noise pollution, which gets to
| extreme levels for somewhat fundamental reasons (you have to
| diaplace a whole lot of air to fly; which is very close to
| having to create sound waves).
|
| So, overall, the energy problem is already fixed, we already
| have point-to-point flying vehicles usable, and occasionally
| used, in urban areas, helicopters. Making them safe when
| operated by a very lightly trained pilot, and silent enough to
| not wake up a neighborhood, are the real issues that will
| persist even if we had mini fusion reactors.
| ynniv wrote:
| I'm not sure that this disproves my original point that self
| driving cars and flying cars don't belong in the same list
| because they are fundamentally different engineering
| problems.
| FabHK wrote:
| Not quite. It's about 3x. It also depends on whether you're
| talking fixed wing or rotary wings.
|
| A modern car might easily have 130 kW or more, and that's what
| a Cessna 172 has (around 180 hp). (Sure, a plane cruises at the
| higher end of that, while a car only uses that much to
| accelerate and cruises at the lower end of the range - still
| not a factor of 10x.)
|
| As another datapoint, a Diamond DA40 does around 28 miles per
| gallon (< 9 litres per 100 km) at 60% power cruise.
| IAmGraydon wrote:
| Does it drive anyone else crazy when an author posts 15,000 words
| (yes, there are that many in this article) when 1,500 would have
| more than communicated the relevant information? The length of
| this article is almost comical.
| andyferris wrote:
| So... he's not a fan of Elon Musk, I take it?
| richrichie wrote:
| > It distorts where VC money goes, always to something that
| promises impossibly large payoffs-it seems it is better to have
| an untested idea that would have an enormous payoff than a tested
| idea which can get to a sustainable business
|
| But this is the whole point of VC investing. It is not normal
| distribution investing.
| kookamamie wrote:
| It's far too rambly and vague to make any sense of the achieved
| results, I think.
| lalaithion wrote:
| > A robot that has any real idea about its own existence, or the
| existence of humans in the way that a six year old understands
| humans
|
| It seems to me we're at the very least close to this, unless you
| hold unproven beliefs about grey matter vs silicon.
| kqr wrote:
| > Systems which do require remote operations assistance to get
| full reliability cut into that economic advantage and have a
| higher burden on their ROI calculations
|
| Technically true but I'm not convinced it matters that much. The
| reason autonomation took over in manufacturing was not that they
| could fire the operator entirely, but that one operator could man
| 8 machines simultaneously instead of just one.
| SavageBeast wrote:
| In reading this I come to wonder if the current advances in "AI"
| are going to follow the Self Driving Car model. Turns out the 80%
| is relatively easy to do, but the remaining 20% to get it right
| is REALLY hard.
| brisky wrote:
| Agree, that is why the agent hype is going to bust. Agent means
| giving AI control. That means critical failure modes and the
| need of human to constantly oversee agent working.
| zetalyrae wrote:
| > LLMs have proved amazing facile with language.
|
| If you took a transcript of a conversation with Claude 3.6
| Sonnet, and sent it back in time even five years ago (just before
| the GPT-3 paper was published), nobody would believe it was real.
| They would say that it was fake, or that it was witchcraft. And
| whoever believed it was real would instantly acknowledge that the
| Turing test had been passed. This refusal to update beliefs on
| new evidence is very tiresome.
| tkgally wrote:
| Similarly if you could let a person from five years ago have a
| spoken conversation with ChatGPT Advanced Voice mode or Gemini
| Live. For me five years ago, the only giveaways that the voice
| on the other end might not be human would have been its
| abilities to answer questions instantaneously about almost any
| subject and to speak many different languages.
|
| The NotebookLM "podcasters" would have been equally convincing
| to me.
| harvodex wrote:
| The whole point of the post is that many have updated their
| beliefs too much.
| coderintherye wrote:
| Good example of everything that can wrong with a prediction
| market if left unchecked. Don't like that Waymo broke your
| prediction? Fine just move your goalposts. Like that prediction
| came true but on the wrong timeframe? Just move the goal posts.
|
| Glad Polymarket (and other related markets) exist so they can put
| actual goal posts in place with mechanisms that require certain
| outcomes in order to finalize on a prediction result.
| sashank_1509 wrote:
| Well said, shows even the most accomplished humans have the
| same biases as the rest of us when not held accountable
| littlestymaar wrote:
| > Glad Polymarket (and other related markets) exist so
|
| Polymarket is a great way to incentive people into making their
| predictions happen, with all clandestine tools at their
| disposal, which is definitely not what you want for your
| society generally.
| HDThoreaun wrote:
| Polymarket suffers from the same problem. This market
| https://polymarket.com/event/ethereum-etf-approved-by-may-31...
| was resolved in an extremely contentious way.
| lifeisstillgood wrote:
| >>> [self driving cars are rmeote controlled] in all cases so far
| deployed, humans monitoring those cars from a remote location,
| and occasionally sending control inputs to the cars.
|
| Wait, What now?
|
| I have never heard this, but from the founder of CSAIL I am going
| to take it as a statement of fact and proof that basically every
| AI company is flat out lying.
|
| I mean the difference between remote piloting a drone that has
| some autonomous flying features (which they do to handle lag etc)
| and remote driving a car is ... semantics?
|
| But yeah it's just moving jobs from one location to another.
| tsimionescu wrote:
| Note that even the examples he gives are related to things like
| an operator telling the car to overtake a stopped truck instead
| of waiting for it to start again. So occasional high level
| decisions, not minute-to-minute or even second-to-second
| interactions like you have when flying a drone.
|
| This is more like telling your units to go somewhere in a video
| game, and they mostly do it right, but occasionally you have to
| take a look and help them because they got stuck in a
| particularly narrow corridor or something.
| RamblingCTO wrote:
| Nitpick: he's not the founder, not by far. He's just a past
| director of CSAIL
| Over2Chars wrote:
| I am always a fan of people who pretend to have psychic powers.
|
| Predict the future, Mr. Brooks!
| barnabyjones wrote:
| >Individually owned cars can go underground onto a pallet and be
| whisked underground to another location in a city at more than
| 100mph.
|
| I'm curious where this idea even came from, not sure who the
| customer would be, it's a little disappointing he doesn't mention
| mag-lev trains in a discussion about future rapid transit. I'd
| much rather ride a smooth mag-lev across town than an underground
| pallet system.
| qwertox wrote:
| Yet such an underground system should exist to transport
| deliveries.
| kweingar wrote:
| Elon Musk promised to build this tech.
| qznc wrote:
| If someone wants to have a credible prediction scorecard, get it
| on some third-party platform like Metaculus, Manifold, GJOpen,
| Polymarket, ...
| maverickmax90 wrote:
| What you marked as hype is a flaw in your skill to recognize real
| world cases vs wishful thinking.
|
| You are not predicting just daydreaming.
| FabHK wrote:
| LOL about the last paragraphs:
|
| > Let's Continue a Noble Tradition!
|
| > The billionaire founders of both Virgin Galactic and Blue
| Origin had faith in the systems they had created. They both
| personally flew on the first operational flights of their sub-
| orbital launch systems. They went way beyond simply talking about
| how great their technology was, they believed in it, and flew in
| it.
|
| > Let's hope this tradition continues. Let's hope the billionaire
| founder/CEO of SpaceX will be onboard the first crewed flight of
| Starship to Mars, and that it happens sooner than I expect. We
| can all cheer for that.
| tontonius wrote:
| How am I supposed to read this ? a thinly veiled hatred for Mr.
| Musk?
| FabHK wrote:
| That's how I interpreted it, yes.
| teractiveodular wrote:
| All that verbiage about robotaxis and not a single mention about
| China, which by all accounts is well ahead of the US in deploying
| them out on the road. (With a distinctly mixed track record, it
| must be said, but still.)
| yearesadpeople wrote:
| Quite an unreadable web page, and somehow rationalising there was
| 'everything before me', and 'everything after me' with regard
| technology and prediction. Unfortunate understanding of reality
| really.
| sinuhe69 wrote:
| I always have a definitional problem with predictions. I mean,
| it's moot whether a _specific_ prediction is right or wrong as
| long as it doesn 't help us to understand the big picture and the
| trends.
|
| Take, for example, the prediction about "robots can autonomously
| navigate all US households". Why all? From the business POV, 80%
| of the market is "all" in a practical sense, and most people will
| consider navigation around the home as "solved" if they can do it
| for the majority of households and with virtually no
| intervention. Hilarious situations will arise that amuse the
| folks; video of clumsy robots will flood the internet instead of
| cats and dogs, but for the business site, it's lucrative enough
| to produce and sell them en masse. Another question of interest
| is how is the trend? What will the approximate cost of such a
| robot be? How many US households will adopt such a robot by which
| time, as they have adopted washing machines and dishwashers. Will
| we see a linear adoption or rather a logistic adoption? These are
| the more interesting questions than just whether I'm right or
| wrong.
| nurettin wrote:
| I don't know the motivation behind making robotics and AI
| predictions, as these things have been done to death since the
| 70s, but I know people who bet for high inflation made a killing
| in financial futures.
| Upvoter33 wrote:
| The next big thing beyond deep learning being LLMs is funny
| michaelbuckbee wrote:
| One interesting prediction category in the Venn diagram overlaps
| of Rodney's predictions that is missing is drone deliveries.
|
| Where I live (in suburbia Virginia), we now can get items from
| the local WalMart grocery via DroneUp, which kind of blows mind.
| sashank_1509 wrote:
| Wow I had no idea progress was being made in this front
| skizm wrote:
| I don't have a pulse on how far self-driving has come from a tech
| standpoint, but from an outsider's perspective I'd say it is
| "achieved" when I can order a self-driving car from an app in all
| of the top 10 most populated cities in the US (since that's where
| it is being developed) with as much consistency as uber/lyft. The
| real final boss for self-driving will be the government red-tape
| that companies will need to get through. I doubt local
| governments will be a laissez faire with self-driving as they
| were with uber being an illegal taxi company.
| metalliqaz wrote:
| the final boss will be the first big lawsuit against a
| manufacturer for liability after someone is killed by a
| driverless car
|
| Of course, then we will eventually see infrastructure become
| _even more_ hostile to non-drivers and people will have to sue
| their own governments for the right to exist in public without
| paying transport companies. Strong Towns tried to warn us
| skizm wrote:
| Didn't that already happen in Phoenix? They paused the
| program there, but unsure if there was actually a lawsuit or
| settlement.
| metalliqaz wrote:
| Wasn't there a human "monitor" in the car?
| mrguyorama wrote:
| There was, but the reality is that the modern US
| regulatory environment demonstrably doesn't care whether
| cars labled as self driving are actually capable of that
| safely, and have not demonstrated any interest in
| regulating such, and that was BEFORE we popularly elected
| a group of charlatans, hacks, and grifters that have all
| made "The courts will bend over backwards for us and our
| wealth" a huge part of America.
|
| Like, Reagan's instructions to the regulatory agencies to
| basically stand down was only just beginning to be undone
| after 40 years, and we immediately elected the people
| promising to slam _hard_ in the other direction.
|
| America will be a regulatory free for all for business
| for decades.
| bArray wrote:
| > The level of hype about AI, Machine Learning and Robotics
| completely distorts people's understanding of reality. It
| distorts where VC money goes, always to something that promises
| impossibly large payoffs-it seems it is better to have an
| untested idea that would have an enormous payoff than a tested
| idea which can get to a sustainable business, but does not change
| the world for ever.
|
| One thing to remember is that there is more than one target
| audience in these claims. VCs for example seem to operate on a
| rough principle of 5 tech companies, 4 make 0x and one makes 10x,
| for a total 2x on each investment. If you only promise 5x, with 4
| failures of 0x and one success at 5x, total return is 1x on each
| (not worth the risk). You may say "yes, my company is 2x, but it
| is guaranteed!" - but they all sell this idea. VCs could be
| infinitely good at predicting success and great companies, but
| it's based on partial information. Essentially companies have to
| promise the 10x and the VCs assume they are likely incorrect
| anyway, in order to balance the risk profile.
|
| I do have a fundamental problem with this "infinite growth" model
| that almost everything seems based on.
|
| > There is steady growth in sales but my prediction of 30% of US
| car sales being electric by 2027 now seems wildly optimistic. We
| need two doublings to get there in three years and the doubling
| rate seems more like one doubling in four to five years.
|
| Even one doubling in 4-5 years might be too much. There are
| fundamental issues to be addressed:
|
| 1. What do we do about crashed EVs? They are dangerous to store
| and dangerous to dismantle. There have been quite a few EV fires
| at places like Copart now. There is little to no value in crashed
| EVs because they are so dangerous, which pushes insurance up
| because they cannot recover these funds.
|
| 2. Most car dealerships in the UK refuse to accept EVs for trade-
| in, because they sit on their forecourt until they eventually
| die. Those who can afford EVs typically get them on finance when
| the batteries provide the fullest range. Nobody I know is buying
| 10 year old EVs with no available replacement batteries.
| Commerical fleets are also not buying any more EVs as they
| essentially get no money back after using them for 3 years or so.
|
| 3. The electrical grid cannot scale to handle EVs. With every
| Western country decarbonising their electrical grid in favour of
| renewable energy, they have zero ability to respond to increased
| load.
|
| The truth is, when they push to remove fossil fuel vehicles, they
| simply want to take your personal transport from you. There is no
| plan for everybody to maintain personal mobility, it'll be a
| privilege reserved for the rich. You'll be priced out and put
| onto public transport, where there will be regular strikes
| because the government is broke and wages cannot increase -
| because who knew, infinite growth is a terrible investment model.
|
| > The other thing that has gotten over hyped in 2024 is humanoids
| robots.
|
| > The visual appearance of a robot makes a promise about what it
| can do and how smart it is.
|
| The real sin is not HRI issues, it's that we simply cannot
| justify them. What job is a humanoid robot supposed to do? Who is
| going to be buying tens of thousands of the first unit? What is
| the killer application? What will a humanoid robot do that it is
| not cheaper/more effective to do with a real human, or cannot be
| done better with a specialised robot?
|
| Anything you can think of which is a humanoid robot performing a
| single physical action repeatedly, is wrong. It would need to be
| a series of tasks that keeps the robot highly busy, and the
| nature of the work needs to be somewhat unpredictable (otherwise
| use a dedicated robot). After all, humans are successful not
| because we do one thing well, but because we do many not-well
| defined things good-enough. This kind of generalisation is
| probably harder than all other AI problems, and likely requires
| massive advances in real-time learning, embodiment and intrinsic
| motivation.
|
| What we need sub-problems for robots, i.e. like a smart vacuum,
| where robots are slowly but surely introduced into complex
| environments where they can safely incrementally improve. Trying
| to crack self-driving 1+ tonne high speed death machines in your
| first attempt is insanity.
| RAHIMUDEEN wrote:
| Big Mumbai
| thefaux wrote:
| > Their imaginations were definitely encourage by exponentialism,
| but in fact all they knew was that when the went from smallish to
| largish networks following the architectural diagram above, the
| performance got much better. So the inherent reasoning was that
| if more made things better then more more would make things more
| better. Alas for them it appears that this is probably not the
| case.
|
| I recommend reading Richard Hamming's "The Art of Science and
| Engineering." Early in the book he presents a simple model of
| knowledge growth that always leads to an s-curve. The trouble is
| that on the left, an s-curve looks exponential. We still don't
| know where we are on the curve with any of the technologies. It
| is very possible we've already passed the exponential growth
| phase with some of these technologies. If so, we will need new
| technologies to move forward to the next s curve.
| ripe wrote:
| On reading the negative commentary here on Rodney Brooks's post,
| I'm realizing that besides being a rambling article, it also
| assumes too much background from the reader. It isn't really
| understandable without knowing something about the author and
| about the business of robots.
|
| Disclaimer: I worked for years building robots, several of these
| years with Rod. I assure you, when it comes to robotics and AI,
| he knows what he's talking about.
|
| Here's my perspective. Also, he wrote his original predictions
| six years ago in a blog post [1], which is the basis for this
| latest post. If you don't have the time to read the old post, I
| provide a short summary from it about autonomous driving below,
| too.
|
| 1. Rod is not just an MIT professor emeritus and a past director
| of CSAIL. He has co-founded multiple robotics companies, one of
| which, iRobot, made loads of money selling tens of millions of
| consumer-grade autonomous robots cleaning floors in people's
| homes.
|
| Making money selling autonomous robots is a very, very difficult
| thing. Roomba was a true milestone. Before then, the only
| civilian, commercially successful mass-produced robots were the
| programmable industrial arms that are still used in auto
| manufacturing. If the author sounds self-important, maybe that's
| why.
|
| Yeah, he can get a little snarky sometimes when self-important
| CEOs run around with VC money in their pockets making tall claims
| and never being held accountable. That's just his style. Try to
| look beyond it. You might learn a thing or two.
|
| 2. The entire purpose of his annual "predictions" posts starting
| with [1] was to counter the hype and salesmanship about AI and
| robotics that's wasting billions of investment dollars and
| polluting the media landscape.
|
| About autonomous cars, he believes that the core technology has
| been demonstrated in the 1980s, but that instead of using it, we
| have squandered the decades since then. For autonomous robots,
| the interaction with their surroundings is critical to success.
| We could have enhanced our road and communications infrastructure
| to enable autonomous cars. Instead, we have chosen to give money
| to slick salesmen to chase the mirage of placing "intelligent"
| cars on existing roads, continuing to neglect our civil
| infrastructure.
|
| [1] https://rodneybrooks.com/my-dated-predictions/
| bhelkey wrote:
| It is valuable to make predictions about the world, evaluate
| those predictions, and reflect on the quality of the predictions
| and what biases skewed those predictions. The key is to refine
| how one looks at the world.
|
| I don't see that in this article. Largely, I see the author
| trying to argue that he was right in 2018 rather than trying to
| take a step back to accurately evaluate their predictions.
___________________________________________________________________
(page generated 2025-01-10 23:01 UTC)