[HN Gopher] The New Inflection
___________________________________________________________________
The New Inflection
Author : break_the_bank
Score : 157 points
Date : 2024-03-19 16:23 UTC (6 hours ago)
(HTM) web link (inflection.ai)
(TXT) w3m dump (inflection.ai)
| fnbr wrote:
| Wow. This is surprising. Not even an acquihire. I'm very curious
| about what happened internally to lead to this.
| riku_iki wrote:
| no more money left..
| tfehring wrote:
| It would be extremely poor form for founders to intentionally
| leave their startup at this stage to take another job. In all
| likelihood they were pushed out by the board.
|
| My guess is that the founders were all in on Pi, their
| conversational language model, and investors lost confidence in
| their ability to build a sustainable venture-scale company on
| the ~4th best AI chatbot.
| GaggiX wrote:
| I've been using their new model for a few days and it works
| surprisingly well, it's also free and works in the EU.
|
| I don't know how I feel about Microsoft's involvement here.
| m_ke wrote:
| No traction, was probably clear they didn't have a shot at
| growing the consumer side.
| GaggiX wrote:
| They have +1 mln daily users.
| azinman2 wrote:
| What's the curve like
| GaggiX wrote:
| +10% every week in the last two months, as reported by
| Mustafa Suleyman.
|
| Edit: sources: https://inflection.ai/inflection-2-5 and
| https://www.axios.com/2024/03/07/inflection-ai-chatgpt-
| opena...
|
| I can not comment more or edit previous comments that are
| not this one, rip.
| riku_iki wrote:
| with capital investments required to keep them afloat, it
| may be not enough.
| DreamGen wrote:
| What's your source on this? They just very recently reached
| 100K downloads on Android and according to various SEO
| tools, they get maybe ~4M _visits_ _per-month_ (and these
| tend to overestimate, plus it 's monthly visits, not DAU).
| rvz wrote:
| It appears that Inflection AI made no sense to begin with and Pi
| was quite frankly a performative research demo and didn't
| generate enough money for the VCs to justify another fundraising
| round. How is Inflection AI worth $4BN?
|
| What can Pi do that is unique over the best of cloud LLMs and the
| hundreds of $0 free LLMs out there?
|
| It appears that it is a vehicle for VCs to quickly run this
| company to the ground for a quick exit, knowing that this company
| is extremely overvalued.
|
| Probably after this acqui-hire, the value of Inflection AI is now
| down to its real value of $200M at most.
| telltruth wrote:
| My guess is that Mustafa wanted to sell to MSFT at 10X but MSFT
| didn't wanted pay that kind of money. Mustafa was ok with fire
| sale but VCs were greedy. Mustafa then quite in rage.
| kcorbitt wrote:
| So sounds like the real news is that Microsoft basically
| acquihired the Inflection founding team?
| nexuist wrote:
| I hope one day we can drop the corporate facade so these blog
| posts can be two to three sentences of "Our company sucked.
| Microsoft wanted to give me a fuck ton of money. I decided to
| take it. Some other guy is in charge now. Good luck!"
| gcnnbdff wrote:
| A lot of people don't know this but all these VC valuations
| are fake, they had this planned years ago to fake growth.
| f6v wrote:
| Maybe someone try training an LLM to do just that.
| bee_rider wrote:
| Does getting acquhired by MS imply your company sucked?
|
| There are things I like about where I'm at but if MS really
| wanted to they could send enough dump trucks full of cash to
| ruin the local traffic situation. Is it worth staying at a
| place if you can't even get out of the driveway to get to
| your favorite little cafe?
| hinkley wrote:
| It means your people are good and your product isn't. That
| requires a kind of cognitive dissonance that some people
| would call "suckage". Certainly had that conversation many,
| many times.
|
| Just be careful you aren't Boeing buying McDonell Douglas,
| where the cut throat culture infected the host and killed
| it, rather than being muted and managed out.
| dbish wrote:
| Eh I think some startup culture injected into Microsoft
| is a good thing, whatever that might mean. Have to shake
| things up
| nyokodo wrote:
| > startup culture injected into Microsoft is a good thing
|
| Companies like Microsoft have consumed countless startups
| and startup employees over the decades and are practiced
| at digesting them without being changed all that much.
| dbish wrote:
| It's a bit different when you're replacing or creating
| the head of a division since they'll replace their own
| lieutenants and onwards down
| RockyMcNuts wrote:
| didn't really get acquihired, which would be if MSFT bought
| inflection, instead it seems MSFT hired away a team,
| probably with options and whatnot.
| aeternum wrote:
| Ironically, AI will probably help to provide this.
|
| Could probably do it now with a browser plugin and the right
| llm prompt. Perhaps call it the un-BSifier.
| OtherShrezzing wrote:
| The prompt "summarise this in 2-3 sentences in the least
| charitable way possible" returned the following on
| chat.mistral.ai
|
| >Inflection is changing its business model again, now
| trying to sell its AI services to commercial clients after
| realizing people can't replicate their AI model. They're
| losing two co-founders to Microsoft and bringing in some
| new CEO to run the show. They claim nothing will change for
| users, but who knows what will happen with all these
| changes going on.
| jack_riminton wrote:
| You could set this up as a service on unbullshitthis.com
| kookamamie wrote:
| Hear, hear! There is a market for a no-bs-happenings.com.
| duxup wrote:
| I think we can assume almost all acquisitions are "this cash
| is better than continuing to operate as we were.... make of
| that what you will".
| guytv wrote:
| I asked ChatGPT to summarize, this is what it came up with:
|
| * Inflection shifts focus to AI studio business, developing
| custom AI models for commercial use.
|
| * Plans to host Inflection-2.5 on Microsoft Azure, with plans
| for other cloud platforms.
|
| * API launch soon, sign-ups for early access open.
|
| * Co-founders Mustafa and Karen to start Microsoft AI,
| leaving Inflection.
|
| * Sean White appointed new CEO; Reid Hoffman remains on
| board.
|
| * No immediate changes to Pi service; privacy and data
| policies unchanged.
| rllearneratwork wrote:
| is it really a win on MSFT part though?
| vineyardmike wrote:
| Absolutely.
|
| Pi (their chat bot) is pretty nice to talk to. It's really
| good with the whole para-social aspect of chatting without
| being weird. Is that useful? Maybe for research, probably
| less so for a direct product.Microsoft can probably get
| something useful from that.
|
| More importantly though, Microsoft is now hosting a whole
| suite of LLMs on Azure. This is the lesson they learned after
| OpenAI had that acute leadership crisis. This is _another_
| hedge against OpenAI. I wouldn't be surprised to see
| Microsoft start distancing themselves from OpenAI in the
| future.
| turnsout wrote:
| Pi is nice to talk to because unlike other chatbots, it
| always, always, always asks a follow-up question. I'd love
| to see the system prompt. That's what's driving their claim
| of high engagement--the bot just keeps drilling you with
| questions, and for some people, the bait is irresistible.
| petre wrote:
| Not only follow up questions, but it also gives straight
| answers without bs, lecturing the user, claiming the
| model is somehow limited or flat out refusals which are
| Gemini's specialty.
| zooq_ai wrote:
| It is not.
|
| Mustafa Suleyman is the ultimate AI grifter.
|
| He has ZERO technical skills. ZERO hardcore STEM background.
| He was just friends with Demis at the right time and road
| that hype train.
|
| At Google he had a similar role VP of AI products or
| something and he contributed nothing (except all the garbage
| ethics, safety crap that didn't help google, probably
| kneecapped it actually)
|
| This is Satya Nadella's worst AI move yet.
| option wrote:
| I am not sure why you are being down voted. Track record
| matters.
| bee_rider wrote:
| AI safety and ethics are a big deal. Nobody wants to buy a
| big heap of linear algebra. A big heap of linear algebra
| that is so close to becoming sentient that we have to plan
| what its ethics should be? Where do I put my money?!?
|
| If you ask engineers to sell linear algebra, they invent
| things like control systems, which are too tricky to sell.
| rllearneratwork wrote:
| "AI safety and ethics are a big deal" - yes. But no
| bigger deal and, arguably, a smaller deal compared to
| software safety and ethics.
| rvz wrote:
| > (except all the garbage ethics, safety crap that didn't
| help google, probably kneecapped it actually)
|
| Yes. Inflection's AI was just too useless and safe to be of
| any use and ultimately made no money.
|
| > This is Satya Nadella's worst AI move yet.
|
| We'll see. But I predict that Inflection AI will eventually
| sunset Pi.
| petre wrote:
| "AI ethics" from a person booted based on allegations of
| coworker bullying?
| telltruth wrote:
| They aquihired 3 people who didn't do any technical work.
| Nadella miscalculated big time here.
| jonny_eh wrote:
| Not if it kneecaps a key competitor.
| verticalscaler wrote:
| That's right, they could be doing no work at Google !
| ipsum2 wrote:
| Fortune reported that Microsoft hired most employees from
| Inflection.
| m_ke wrote:
| Right, Karen Simonyan is not technical at all...
| telltruth wrote:
| They had raised massive amount and not from good patient
| investors. No traction means Mustafa got fired. This is not
| surprising though but what is surprising is MSFT picked him up.
| The guy is not technical, is not even visionary and had just got
| lucky hanging out with Demis. I would think Satya had better
| taste.
| toomuchtodo wrote:
| Never underestimate the power of soft/social skills.
| rqtwteye wrote:
| That's where the real money is. Even the best technologists
| won't make it very far if they don't know how to play
| politics.
| skepticATX wrote:
| Mustafa wasn't exactly on his best behavior at DeepMind, if the
| allegations are to be believed. It really is surprising that
| Microsoft would hire him.
| refulgentis wrote:
| It's crazy there's zero accountability for bad behavior in
| tech. I went through my own story at Google, and seeing them
| say it was vaguely bad before promoting him to VP mirrors
| exactly the "intervention" I saw.
|
| The deck is completely stacked against you based on
| hierarchy. Behavior that a fast food manager would
| proactively solve in 30 seconds gets ignored in white collar
| tech. No one above you will even mention it - they know you
| can't win and they just hope you'll quietly give up.
|
| If someone above you in the informal hierarchy is messing
| with you, there's massive confirmation bias if you complain.
| They'll spin it to whoever you complain to make you the bad
| guy. HR never helps - their job is to investigate, and then
| give the results to someone 2-3 steps above you to do
| something with.
|
| The higher ups control the outcome, and they designed the
| power structure in the first place, their confirmation bias
| is accept the spin.
|
| If you want to survive, avoid conflict 100% of the time. Let
| people blame you, fail reviews undeservedly.
|
| My Google career ended from just doing exactly what I was
| supposed to do in order to get a 3 year delayed project done,
| that 4 separate VPs had been asking for all those years. I
| spent 6 months warning my manager fuckery was afoot. Didn't
| matter. TPM witnessed and defended me, didn't matter. Guy who
| led it hired his unqualified childhood buddy to replace me.
| Didn't matter. All on me. Everyone wanted to do it, and gee
| whillakers, refulgentis went mad and dropped the ball
| completely for some reason.
|
| Of course, 6 months later they delayed the project a 4th year
| because they could, documenting the only downside being a
| strained relationship with a less influential partner team.
| (my orgs managers didn't realize their...unvarnished...takes
| were in a doc shared with all of Google)
|
| At the end of the day, HR will funnel you into taking mental
| health leave -- 6 months worth, exactly long enough that an
| EEOC complaint can no longer be filed. (took me 6 years to
| realize why "disgruntled Google employee" news articles
| always included a bit referencing leave/6 months off as if it
| was a bad thing. go/mh-leave if you're at Google. You don't
| actually need to talk to HR, and I don't recommend going to
| them ever. I didn't for this, but they wouldn't have helped.)
|
| The whole system is broken.
| VirusNewbie wrote:
| >Guy who led it hired his unqualified childhood buddy to
| replace me. Didn't matter. All on me. Everyone wanted to do
| it, and gee whillakers, refulgentis went mad and dropped
| the ball completely for some reason.
|
| How is that possible at Google, which should have a hiring
| committee? Managers aren't allowed to just hire rando
| person.
|
| Did your director or VP not like you or something? I'm
| curious if there's more to this story.
| refulgentis wrote:
| There always is, things that mattered here:
|
| Why didn't higher ups care? I didn't bother going up
| higher than skip. My core interlocutor was my skip's
| peer's report's report, I didn't expect my skip to go to
| war over slow-drip white collar bullying. Honestly, I was
| done and planning my exit once year 3.5 of "not this
| year" hit, going crying to VPs you see once a month /
| once a quarter felt insane & would have just devolved to
| he said/she said.
|
| To your point re: seems like a lot. I worked with a
| couple counselors at same level as my skip over my last
| year there. (highly recommend G2G if anyone reading is at
| Google, kept me sane.) #1 said they dealt with less after
| kissing their VP's wife at an off-site - which is why I
| got #2, wasn't sure if that one was too skeezy at first.
|
| What would I have done differently? I was honest the
| whole time, which didn't help because the fact I wasn't
| happy and it was escalating was clear. ex. with the
| hiring friend thing, told my manager that I was shocked
| and didn't expect that kind of thing at a startup.
|
| How do you hire a friend without domain experience as a
| manager?
|
| There's a core principle that once you've made it through
| Google interviews, domain doesn't matter, all Google SWEs
| will excel. Having loose rules with kind intent is
| awesome, but they're double-edged. That gives you
| rationale, and combined with moving recruiters into
| individual orgs., lets you put your thumb on the scale
| and bring in who you want. Also 2022 through June, Google
| was _desperate_ to hire, this would have been justified
| as an awesome referral, and what 's a friend, anyway?
| Bjorkbat wrote:
| He also left DeepMind because of allegations of bullying
| employees (https://en.wikipedia.org/wiki/Mustafa_Suleyman).
| Between that, what you just brought up, and the strange PR
| blitz he went on to promote his book, I kind of predicted
| Inflection would run into major trouble well before OpenAI and
| Anthropic
| Workaccount2 wrote:
| Some of the most damaging people I have met are people that you
| are absolutely enthralled with when you first get to know them.
| spaceman_2020 wrote:
| True, and some of the most effective people I know haven't
| been impressive when you first meet them.
|
| I'm now extremely wary of blusteringly confident, glib
| talkers. Experienced a high correlation between these traits
| and sociopathy.
| geraldhh wrote:
| quite sure it's just unlucky ppl that got rewarded for
| acting like this
| rqtwteye wrote:
| That would apply probably to most famous entrepreneurs.
| Whatever I have read about Jobs, Musk, Gates and others, they
| all are very willing to abuse people to achieve their goals.
| maest wrote:
| A => B <> B => A though
| coffeebeqn wrote:
| First impressions is a sales skill
| baq wrote:
| There's a whole extremely famous book series on this very
| topic, you might have heard about it. The first volume's
| title is 'Paul is bad' but it's more widely known as 'Dune'.
| DaiPlusPlus wrote:
| Dune is about worms.
| kevindamm wrote:
| Dune is about why you don't mix politics and religion.
| vouaobrasil wrote:
| > As part of this, we're thrilled to announce that we will now
| host Inflection-2.5 on Microsoft Azure helping us get it into the
| hands of creators everywhere.
|
| Hahah. You know, the word 'creator' has truly become co-opted by
| consumerism and greedy bigcorps. What a lot of posts like this
| never mention is one of the primary uses of AI: to make more
| efficient the system of trying to get us to buy even more things
| we don't need. I mean, who are we kidding? In a well-balanced
| life, we shouldn't need or even interact with personal AIs. We
| should slow down and appreciate what we have and seek for
| simplicity.
|
| If there's too much data to be handled in business, it means
| business is not going in the right direction, not that we need
| new tools to handle it.
|
| If we are pressured to write more, then we are creating things
| not of true value but merely to amuse.
|
| If we feel like becoming more efficient is a good thing, it
| simply is an extension of the original psychological manipulation
| of advertising demanding MORE for the industrial machine.
|
| Let's not fool ourselves into thinking this garbage is something
| good.
| airstrike wrote:
| It's hard to agree with most of what you said without starting
| from the big assumption that "capitalism is bad" and "being
| anti-capitalism is a virtue"
|
| Becoming more effective in and of itself simply means becoming
| more productive, which means creating goods and services with
| less, which means more "welfare" to people
|
| If people are being convinced to buy shit they don't need,
| that's a separate issue. But don't throw the baby out with the
| bathwater
|
| In _my_ well balanced life, I love interacting with LLMs. I
| also appreciate the things I have all of the time, without
| needing to slowing down technological progress to do so
| vouaobrasil wrote:
| > In my well balanced life, I love interacting with LLMs. I
| also appreciate the things I have all of the time, without
| needing to slowing down technological progress to do so
|
| Okay that's fine, but you are not the rest of the world.
| While you enjoy the ultimate fruits of capitalism, there are
| people in developing countries suffering because of it
| through a lack of freedom and a destructiom of the landbases
| that they could once depend on. When climate change comes
| because of consumerism, it is you and other well-off people
| that will be able to move first.
|
| OF COURSE, you will disagree with me, because you are
| probably at the apex. (Just the very fact that you have a
| well-balanced life means you are close to it, proportionally
| speaking to the entire human population.)
|
| I don't claim that all capitalism is bad BTW. Rather, what I
| claim to be bad is global capitalism where other
| considerations have been eliminated or decimated.
| GaggiX wrote:
| I don't think I've ever seen a more cliched argument
| against LLMs than "some people in the world suffer from
| capitalism".
| vouaobrasil wrote:
| That is not my argument against LLMs. Please don't put
| words in my mouth. My actual argument against LLMs is
| that:
|
| 1. They will replace so many jobs en masse that they will
| cause a large loss of meaning in many lives, that they
| will concentrate wealth in the hands of the tech
| companies, that they will enable even more efficient
| propaganda, and that they will be used as a drug to
| distract the populace from increasingly destructive
| practices enacted by global capitalism
|
| 2. LLMs will isolate us by providing more and more
| services that once were provided by humans. AI
| therapists, girlfriends/boyfriends, solvers of problems
| that we once turned to our fellow human being to solve.
| This vast increase in individual autonomy, pushed by
| psychological manipuluation to THINK that we need such
| autonomy, will destroy more local communities and
| connections that people have with each other. In
| conclusion we will become isolated and miserable, relying
| completely on the machine to keep our lives balanced
| without any autonomy to seek out our own way of doing
| things.
|
| Is that still cliche enough for you, or would you like to
| make more comments in bad faith about my position?
| GaggiX wrote:
| >That is not my argument against LLMs. Please don't put
| words in my mouth.
|
| That was the most people could read thought from your
| previous comments, now that I make you notice that, you
| wrote some actual arguments, even though I do not share
| your doomerish vision of the future.
| drivebyhooting wrote:
| Don't put thoughts in my head. I could understand his
| comments perfectly.
| yoyohello13 wrote:
| I think you are absolutely correct in your assessment and
| I think there is absolutely nothing that can be done
| about it.
| airstrike wrote:
| People in developing countries are better off with
| capitalism than without it... speaking as someone born and
| raised in a developing country. No, capitalism isn't
| perfect, but then again nothing is
|
| Speaking of Brazil specifically since that's in your
| nickname, the deforestation of the Amazon to make room for
| cattle isn't caused by developed countries running LLMs.
| It's caused by Brazilians who are negligent, complacent,
| happy, even, with that choice.
| geraldhh wrote:
| true, capitalism is causing deforestation. pretty bad
| already, no llm needed.
| ggpsv wrote:
| I find it curious that your reaction to this is to reduce
| OP's point to the "capitalism is bad" trope when it is in
| fact more nuanced than that. His point isn't anti-capitalist
| per se - it critiques the traits of the dominant economic
| system that many of us currently live in.
|
| I won't assume bad faith on your part, but your argument also
| relies on a big assumption with regards to productivity
| gains, and who reaps them, and at the detriment of whom and
| what.
| airstrike wrote:
| Well, up until 5 minutes ago his bio explicitly mentioned
| "anti-capitalism" so I put 2 and 2 together
| vouaobrasil wrote:
| I changed it because anti-capitalist is too vague. I have
| to think of a better term.....
| BriggyDwiggs42 wrote:
| Pretty hard agree here, but I think there's space for a gripe.
|
| Our default state as humans is pretty awful and materially
| deprived, so I think making technology to improve our
| conditions is good. Making technology that does powerful things
| in the real world is difficult partly because of the vast
| quantity of data in its structure. If we make powerful AI
| systems equipped to handle that scale and built to make our
| lives better, it would be good, but we're failing to do that
| which is the issue.
| vouaobrasil wrote:
| > Our default state as humans is pretty awful and materially
| deprived, so I think making technology to improve our
| conditions is good.
|
| It's not a dichotomy! When I rail against technology, I am
| not saying that everyone need go back to the stone age. But
| don't you think there might be a happy medium?
|
| At least personally, I've found that giving up a fair amount
| of technology in many cases has actually improved life in
| ways that I didn't think possible. Not everyone needs to be
| ascetic, but we seem to be headed to a life very tightly
| integrated with technology, and I think it's arguable that
| THIS POINT is past the point of that happy medium.
| ggpsv wrote:
| "It is difficult to get a man to understand something when
| his salary depends upon his not understanding it".
|
| This quote summarizes much of the discourse in HN and the
| tech industry at large.
|
| As you point out, it is in fact a false dichotomy but it is
| hard to run counter to hundreds of years of cultural
| conditioning that has run amok since the start of the
| industrial revolution. Hobbesian thought is common in these
| threads!
|
| Luckily, we've had recent thinkers like Ellul, E.F.
| Schumacher, and Ivan Illich to challenge these entrenched
| views.
| geraldhh wrote:
| The Shirky Principle is not to be underestimated, yet
| more of a humorous take imho.
|
| thou i would question the change in 'cultural
| conditioning' brought by industrialization als in
| practice it's mostly a means to scale society.
|
| schumacher, being an atheist, is probably the most
| agreeable of the authors you mentioned.
| tazu wrote:
| I've started thinking technology is best when it's
| invisible.
| BriggyDwiggs42 wrote:
| I think much of our technology today is just about
| consuming more crap, and all that could go and we'd be
| fine, but I also think that our, say, medical technology
| could be infinitely better than it is. There's also cool
| potentials like space travel or intelligence augmentation
| that I wouldn't want to concede.
| ChildOfChaos wrote:
| Wouldn't a personal AI help you live a more well-balanced life,
| allowing us to offload tasks to it so we can appreciate the
| world around us and simplicity, rather having to do it all
| ourselves?
| vouaobrasil wrote:
| So far it seems that other automations have not done so.
| We've offloaded many tasks to computers, but now we spend
| more time than ever in front of a screen (see
| https://www.statista.com/statistics/645644/north-america-
| dai... for example). More people live in cities, with less
| opportunities for fresh air. And let's not forget that the
| system itself is so complex now with so many dependencies
| that we surround ourselves with anything but simplicity.
|
| There might be a world where what you say is possible of
| course, but I don't see it in this one when it comes to more
| advanced forms of automation.
|
| In fact, sometimes the joys of life can be found in the most
| simple of mechanical tasks. Of course, not everyone thinks
| so.
| crooked-v wrote:
| The problem is that none of them are actually good enough for
| that yet. The closest I've seen are calendar-managing LLM
| services, and even those only really work because it's a
| tightly-constrained task set and easy to confirm the results
| as good or not.
|
| Edit: I still can't even trust Siri to be fully accurate 100%
| of the time about turning the lights on and off, let alone
| anything more complicated than that.
| jjtheblunt wrote:
| Why not just simplify, and not need such?
| Aerbil313 wrote:
| The only inflection I see here is Microsoft single-handedly
| dominating global tech industry and its future after its years in
| sleep. US DoD must have been really worried about the US Empire's
| continued dominance.
| unethical_ban wrote:
| Uh-huh.
|
| Amazon, Google, Meta, and Apple are has-beens.
|
| A plausible theory with less coordination required: Microsoft
| is a large company with lots of money attempting to regain
| competitive advantages against their rivals through
| acquisition.
| CuriouslyC wrote:
| While that's a true story today, Satya is brutally outplaying
| Google, Amazon is a utility company and Apple's domain is
| hardware and will likely continue to be. Mark is making some
| good strategic plays to keep the field even, but Microsoft
| really is poised to clean up as AI takes off.
| danpalmer wrote:
| Microsoft is poised to do well, sure, but they're currently
| beholden to Nvidia for hardware and OpenAI for models, and
| their main surface area for products is Office, Azure, and
| to a lesser extent, Outlook.
|
| Comparatively, Google controls their hardware, Google
| builds their models, and Google has Search, Workspace,
| Gmail, Android, Cloud, and YouTube, all of which could have
| significant AI investments.
|
| I'm not saying Satya isn't doing a great job, he really is,
| and I have many criticisms of Google's approach, but I do
| think Google is at least equally well positioned, and
| possibly better positioned.
| CuriouslyC wrote:
| I agree that TPUs are a huge competitive edge. That and
| huggingface integration is going to let google do very
| well serving open models until other competitive hardware
| comes online. The problem is that I think the window to
| capitalize on TPUs isn't going to be long enough to
| cement any sort of dominance, and they're fumbling so
| badly now that I don't have faith in the leadership.
| bugglebeetle wrote:
| Between this and the Mistral deal (that's currently under
| investigation by the EU), Microsoft looks to be really trying to
| get back into the monopoly business with AI.
| baq wrote:
| he who controls the spaice, controls the universe, after all.
| thomastjeffery wrote:
| Microsoft has been in the monopoly business since at least
| 2001.
| darth_avocado wrote:
| Is tech hiring great? No. Is it the absolute worse it can be?
| Also no. Internship or apprenticeship instead of tech interviews
| are just take home tests that last multiple months, and that is a
| huge waste of time.
|
| Yes, broken interview system currently rewards people who
| interview more, but when you combine it with other signals like
| resume, project presentations and other things, the chances of
| hiring bad candidates are low. Yes, you miss out on a lot of
| great candidates, but that is not a problem that big tech
| companies need to solve. Besides, not knowing the status of your
| employment 3-6 months down the line is not a great thing for
| candidates.
| Analog24 wrote:
| Internships are definitely not a waste of time. First of all,
| they pay well in and of themselves (at least in tech). Second
| of all, most internships are filled by students on their summer
| break. What better use of that time than getting an inside view
| of a company they might want to work for? From the companies
| perspective it gives them a much better idea of how well the
| candidate performs and how they fit in at the company, giving
| much higher confidence in a hiring decision that could lead to
| significant future impact.
| darth_avocado wrote:
| I think you misunderstood. Internships for students during
| the summer definitely makes sense. However, i was talking
| about internships/probation as a way to evaluate candidates
| instead of interviews. Meaning if you are a staff engineer
| with 10+ years of experience, you're still going to be hired
| conditionally for 3-6 months to see if you're good enough for
| the team. I personally find that very prone to abuse.
| bcopa wrote:
| Pretty mind-boggling how ChatGPT wrapper apps built by dev houses
| in Poland are absolutely crushing it, while Pi failed to gain any
| traction despite raising billions.
| BoorishBears wrote:
| More mind-boggling they raised billions to do something any LLM
| can do with some prompts (be more personable)
| Pandabob wrote:
| Have not heard of these. Where can I learn more about them?
| toisanji wrote:
| examples please
| dbish wrote:
| Which?
| lumost wrote:
| The interesting emphasis to me is commercial customers. They are
| acknowledging that the competition to be the nth gpt vendor is
| too stiff for them, and they aren't successful as an independent
| venture funded research lab.
|
| I can think of 4-10 other large vc funded operations in this
| boat.
| lvl102 wrote:
| This space is moving so fast and it's hard to keep track of SotA.
| How are management even keeping track? I suppose MSFT has enough
| cash to buy 'em all.
| coffeebeqn wrote:
| State of the art is still GPT-4? Others are playing catch up or
| hitting very similar benchmarks
| andy99 wrote:
| Is Microsoft AI the thing that Altman was going to be in charge
| of in the short lived agreement he had with MS?
| mupuff1234 wrote:
| Wasn't Mustafa kicked out of deepmind/Google due to harassment
| allegations? Seems like a pretty pathetic move from MSFT.
| Animats wrote:
| Ah, PR. Buzzword buzzword buzzword ... second screen ... "two of
| our three co-founders ... will be leaving Inflection to start
| Microsoft AI",
| gwern wrote:
| Who gets the GPU clusters...?
| ShamelessC wrote:
| Do they own their GPU clusters?
| nabakin wrote:
| I have some serious issues with this company's PR.
|
| They say their Inflection-2.5 model is the world's best personal
| AI[0] which is a dumb claim to make considering it is done off of
| automated benchmarks which we know are flawed and even if you
| assume automated benchmarks are good enough to claim that title,
| it would be held by dozens of other open weight models on HF, not
| Inflection-2.5.
|
| They said their Inflection-2 model was the second best in the
| world[1] while comparing it to Palm-2 which no one considers
| close to the best in the world. They again, based their claim on
| automated benchmarks which anyone knowledgeable in the space
| would know can be gamed and is not representative of actual
| conversational performance. (take a look at the Lmsys Arena
| Leaderboard for a better metric)
|
| They list other models they consider good while failing to
| mention or compare to Mixtral 8x7b, the best open model that
| exists.
|
| And they introduce buzzwords no one in the area uses like IQ and
| EQ as if they are innovative concepts.
|
| Making big, bold claims without evidence is the exact kind of
| manipulative PR speak I'd expect from a company with little to no
| substance.
|
| [0] https://inflection.ai/inflection-2-5
|
| [1] https://inflection.ai/inflection-2
| posix86 wrote:
| Agreed, read a few sentence & move on.
| maest wrote:
| Fwiw I wasn't familiar with them and found gp's comment
| useful. I actually upvoted
| nabakin wrote:
| I think it's good to raise awareness of bad practices if you
| recognize them.
| Leary wrote:
| Exactly, honestly I think their business model could work if
| they have a GPT-4 level model for subscription.
| RaftPeople wrote:
| "After just 3 days of being incorporated and just 2 days of
| actual work performed, we have now created the best AI on the
| planet with trillions of active subscribers, per second.
|
| The quality of our product is so high, it is being compared to
| a perfect diamond the size of the entire solar system. But, we
| are humble, and while we think our tech is pretty darn good, we
| know that we can do better. That's why we are introducing v2 in
| just 2 business days from today, which will render all other
| forms of intelligence (human and artificial) irrelevant."
| fullstackchris wrote:
| Tell me about it. I'm really starting to get ticked off trying
| to wade through garbage hype and actual improvements across
| everything "AI"
| bevekspldnw wrote:
| I just skimmed and came away with "Satya is dumping more Azure
| credits on some AI thing to screw with Big G again". The rest
| was noise.
| petre wrote:
| Dunno, it's pretty damn good and quite useful. I asked Mistral
| 7b and Claude Opus to count the banned substances on the WADA
| anti doping list. Both avoided a straight response giving back
| loads of bs. Claude was next to useless. Mistral gave a
| somewhat acceptable answer counting substances on every chapter
| after much persuasion. Inflection's pi AI gave me a straight
| answer to the first question without any bs. I like its casual
| tone and the fact that it does not use needlessly superfluous
| language like the GPT models.
| crooked-v wrote:
| Also, and I'm going to put this real bluntly, a "personal AI"
| isn't worth crap to a lot of people until it's allowed to talk
| about sex.
|
| I'm not even talking for porn purposes. Sex is a basic and
| healthy part of most peoples' lives, any number of relationship
| issues revolve around it, and for plenty of people it's
| literally their profession, legal or not. Anything that
| pretends it can 'personally' help a broad spectrum of people
| while treating sex as verboten is just bullshit.
| ChrisArchitect wrote:
| Related MS blog post:
|
| _Mustafa Suleyman, DeepMind and Inflection Co-founder, joins
| Microsoft to lead Copilot_
|
| https://news.ycombinator.com/item?id=39757330
| ashvardanian wrote:
| I've said it before, and I'll say it again. The Silicon Valley TV
| series desperately needs a reboot.
| joshmarinacci wrote:
| I don't know about Inflection, but their new CEO is a great
| leader who I've worked with in the past. Just having him on board
| raises their status significantly in my mind.
| kazinator wrote:
| This kind of PR pep talk is not a good place to go on a tangent
| about how three of your co-founders bolted to Microsoft do do AI
| stuff there. #lolwtf
| intellectronica wrote:
| Could this be related to the recent news that Apple is in
| advanced stages of choosing Google as its AI provider? Maybe the
| last hope for Inflection was to become Apple's AI provider and
| when that fell through it was time to put an end to the misery.
| vineyardmike wrote:
| This has only helped their small odds of becoming Apple's
| provider.
|
| You can't serve the LLM for every iOS device in the world
| without a big pile of inference machines and the only companies
| that fit that bill are Microsoft, Google, and Meta.
| syntaxfree wrote:
| I talked to Pi for one New York minute and it came up with the
| good ole "Ahhh the X Y of Z" LLM-ism. So much for an unique
| conversational tone.
| htrp wrote:
| Inflection on deathwatch now?
___________________________________________________________________
(page generated 2024-03-19 23:00 UTC)