[HN Gopher] TL;DW: Too Long; Didn't Watch Distill YouTube Videos...
___________________________________________________________________
TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the
Relevant Information
Author : pkaeding
Score : 234 points
Date : 2025-02-12 02:15 UTC (3 days ago)
(HTM) web link (tldw.tube)
(TXT) w3m dump (tldw.tube)
| gus_massa wrote:
| I tried with a few Thunderf00t videos. He has good analysis, but
| the guys repeats everything too many times. Many are about silly
| impossible "inventions" / scams, but this is an experiment that
| he published in Nature Chemistry:
|
| https://tldw.tube/?v=LmlAYnFF_s8 " _High speed camera reveals why
| sodium explodes!_ -- > " _Coulombic explosion. (Sodium and water
| reaction)_ "
| trainsarebetter wrote:
| I used to enjoy watching his stuff, but he became hyperbolic,
| and now has too much rage bait for the algorithm.
| dvh wrote:
| That was the first thing that came to my mind, all his 20
| minutes videos contains at best 2 minutes of information.
| ryanmcbride wrote:
| I thought maybe I'd finally be able to through a Wendigoon video
| with this, since I'm generally interested in the things he talks
| about but can't stand some of his linguistic tics. Unfortunately
| looks like most of his videos are too long for this, which I
| guess is ironic considering the name.
|
| TL;DP
| ldenoue wrote:
| Here's one video above 1 hour and it works with Scribe
| https://www.appblit.com/scribe?v=FQUo2r-ow-k
| Ancalagon wrote:
| Can't wait till futurama style streaming-into-my-brain is
| available
| zorgmonkey wrote:
| I tried a couple videos in both this site and Kagi's summarizer,
| both were decent but each time Kagi did better.
| mtlynch wrote:
| Whoa, I've been using Kagi for two years and didn't realize the
| summarizer could do videos!
| freeAgent wrote:
| It only works when there's a transcript. It's not "watching"
| the video. That said, it works very well most of the time for
| me.
| rapind wrote:
| This feels like the inevitable outcome of the youtube algorithm
| favouring longer videos.
| swyx wrote:
| does it? i get pushed so many shorts and 15 minute video essays
| graeme wrote:
| 15 minutes actually is longer than what early videos were.
| The fact that it seems short is a reflection of the
| algorithmic push for longness.
|
| Shorts are a whole different thing. But there's a void of
| videos from 90 seconds to 7 minutes that just answer a
| question.
| kragen wrote:
| Yeah, for many years it didn't allow uploading videos of
| over 10'.
| RockRobotRock wrote:
| How can anyone make a generalized statement about "the
| algorithm" when it is by definition personalized?
|
| All YouTube cares about is keeping you on the platform as long
| as possible.
| greenchair wrote:
| because there are generalized algorithms in use as well. The
| 'show ad' logic that is used at specific points along the
| way.
| garbagepatch wrote:
| It would be nice if it also gave the videos better titles. Then
| make it an extension to replace the titles automatically.
| zonkerdonker wrote:
| Ask and ye shall receive. Not only titles, but thumbnails as
| well!
|
| https://dearrow.ajay.app/
| energy123 wrote:
| For specific questions I get the transcript from
| https://tactiq.io/tools/youtube-transcript then copy paste to a
| LLM and append my question
| asah wrote:
| Haha maybe not.
|
| https://tldw.tube/?v=A4kpVO56OBU
| kragen wrote:
| What I'd really like is a service that edits down YouTube videos
| by removing all the stock footage and talking head crap, then
| speeding up the audio to fit over the remaining novel information
| --whether that's new battlefield footage, electron micrographs,
| demonstrations of machining techniques, or just elephant
| toothpaste. The talking head filler seems like it would be easy
| to recognize, but stock footage recognition presumably would have
| a significant false negative rate, which is okay.
|
| This would reduce some videos to just a transcription, which
| would be the ideal outcome, I think. The less of my limited time
| on Earth I waste watching some dumbshit reading a script at a
| camera, the better. Summarizing the transcript further like this
| site does might be occasionally useful, of course.
| satvikpendem wrote:
| I use SponsorBlock (be sure to enable all categories of
| blocking such as filler content, not just sponsors), DeArrow
| (de-clickbaits thumbnails and titles) and a video speed changer
| extension to enable much of your stated functionality, though
| of course not all. I've saved likely years of watching due to
| this combination.
| soraminazuki wrote:
| How did your viewing experience change after using DeArrow?
| I'm curious because DeArrow has a great reputation, yet my
| gut feeling tells me that I should avoid watching clickbait
| videos altogether.
| satvikpendem wrote:
| Then you'd be in luck as it makes me less likely to click
| on those videos. I only watch them if I'm truly interested
| in the topic as shown by the revised thumbnail and title,
| which makes my viewing quality much higher on average.
| eco wrote:
| I love the DeArrow experience. The problem with clickbait
| on YouTube is that even good creators are compelled to use
| it because it works. Whether it's due to the algorithm or
| just human behaviour I don't know.
|
| There's a little blue button next to the video titles to
| toggle it on and off and it's very interesting to compare
| the community title and thumbnail with the original. Good
| original titles are often left largely intact, maybe with
| some extra context. Betteridge law is no longer relevant.
| It's nice.
| pinoy420 wrote:
| Tom Scott and LTT did research on this and you are
| _absolutely correct_. It just "works" for engagement.
| Even those who are already your top viewers.
| Timwi wrote:
| DeArrow lets me make an informed decision on whether to
| watch a video or not. It's true that sometimes the
| clickbait will extend into the first few seconds of the
| video, but you can still stop watching then if you think
| it's not for you. Myself, I enjoy knowing what it's really
| going to be about.
| IshKebab wrote:
| I tried DeArrow for a bit but I found that for some
| channels it actually made things worse - I mean it removed
| the clickbait, but _some_ clickbait gave me a better idea
| of what the video was actually about than DeArrow 's boring
| titles.
|
| Overall it didn't improve the experience really.
| sebzim4500 wrote:
| Just looking at their own marketing material, in some cases
| it's just making the titles longer and more boring without
| actually making them more descriptive.
| eco wrote:
| One of the best things in SponsorBlock is Highlight segments.
| If the video is 10 minutes of filler building up to a single
| interesting/exciting moment you can often see exactly on the
| timeline where to jump to.
| lolinder wrote:
| That's also present in the regular YouTube video scrubber.
| Maybe less prominently?
| _Algernon_ wrote:
| When you need 4 extensions (counting ublock as well) to make
| a site useable, maybe it is time to reconsider what site you
| spend your time on?
| pinoy420 wrote:
| What a silly point of view. Instead what should they do?
| Post comments like this on HN?
|
| They clearly enjoy watching and digesting information from
| YouTube- the biggest platform for this. But people need to
| make money from it hence the ailment these these plugins
| remedy
| _Algernon_ wrote:
| >Instead what should they do?
|
| Read a book, touch grass, etc.
|
| >They clearly enjoy watching and digesting information
| from YouTube- the biggest platform for this.
|
| I'd probably enjoy heroin but I still choose not do take
| it because I know it's not good for me. What's your
| point?
| exe34 wrote:
| > Read a book
|
| I wanted to say something flippant, but dang with the
| halo on my shoulder urges me otherwise.
|
| so here goes: books can also full of filler crap and
| useless and outright wrong material. just because it's a
| different medium doesn't make it any better for wasting
| time. in fact, a bad book can waste weeks of your life
| whereas a bad video on YouTube could waste about 10
| minutes.
|
| ultimately I approve of any tools that allow one to
| extract the information that they are after without
| paying the cost of giving more attention to somebody than
| they've earned.
| _Algernon_ wrote:
| Books at least go through a minimal amount of vetting
| before they are published, and there are meaningful
| differences between forms of media. As McLuhan wrote:
| "The medium is the message".
| Tuna-Fish wrote:
| Lol, no they don't.
|
| If you have a pdf with enough words in it, you can go
| from that to a published paperback on Amazon in 15
| minutes.
| _Algernon_ wrote:
| Those aren't books. They're printed blogs you pay for.
|
| Every book can be published as a pdf. That doesn't mean
| that every pdf is a book.
| satvikpendem wrote:
| What a textbook (sorry, printed blog) example of a No
| True Scotsman.
| shkkmo wrote:
| > Those aren't books. They're printed blogs you pay for.
|
| If you are going to make an outlandish claim like this,
| at least try to post some criteria to defend. Not doing
| so has led to you not realizing that you can't without
| also excluding works of classical literature like Sense &
| Sensibility (which was self-published by Jane Austin).
|
| There's also more direct problems with the arbitary line
| you are trying to draw, like best-selling book The
| Martian, which was originally published as a blog, then
| self-published as an e-book, then officially published as
| a book by a major publisher.
| smsm42 wrote:
| No true Scottsman!
| _Algernon_ wrote:
| Actually its implication:
| https://en.m.wikipedia.org/wiki/Implication
| abecedarius wrote:
| 1. Some blogs are better than most books.
|
| 2. Some carefully-written ebooks now are self-published,
| and I don't care if their cover design and promotion
| campaign aren't as professional.
| IshKebab wrote:
| Maybe in the 1900s. Not any more
| exe34 wrote:
| so if they're good, they're books. if not, they're not
| _really_ books.
| kragen wrote:
| Being a different medium does in fact make it much better
| for wasting time. Skipping a paragraph in a book takes
| 100 milliseconds. Referring back to a previous paragraph
| to understand a seeming contradiction does too. Pausing a
| book requires no effort, and no effort to return to the
| place where you were. Books can be excerpted and quoted,
| not just in other books, but also in conversation or in
| videos, in a way that videos just can't. (Talking-head
| filler scripts can be quoted that way, but the video can
| only be quoted in video or still images.) Errata sheets,
| errata websites, and later editions of books can correct
| errors in earlier editions. Books are much smaller and
| therefore easier to archive than videos, making them less
| likely to get lost.
|
| Neil Postman's _Amusing Ourselves To Death_ is an
| extended and very persuasive reflection on precisely this
| difference in the tendencies of these two media, although
| particularly attuned to the form of video that was
| popular at the time, TV, which had commercials and didn
| 't even have pause and rewind.
| exe34 wrote:
| somehow I doubt that the person I was replying to would
| consider ebooks as a book.
|
| videos can be viewed in double speed, and they can be
| rewound. they can be watched over and over if needed.
| they can describe/demonstrate information in ways that
| books can only project into 2 dimensions. a picture is
| worth a thousand words - well imagine 24 of those per
| second.
|
| certainly with STEM subjects, nothing beats a good
| animation.
|
| even for things like history/geography, I find animations
| and progressive disclosure to be much more engaging and
| easier to remember than dry words on a page.
| pinoy420 wrote:
| "Touch grass"
|
| This turn of phrase says everything to be honest.
| jasode wrote:
| _> Read a book, touch grass, etc._
|
| There's a ton of useful information that's only available
| in Youtube videos and not books such as DIY tutorials,
| repairs, etc.
|
| Unfortunately, many helpful videos also include useless
| fluff. E.g. a "good" 10 minute video on how to
| disassemble a Samsung dryer to replace the heating
| element also has 2 extra minutes of _" click my link to
| Nord VPN"_ and _" please smash that Like button and
| subscribe"_. Therefore, tools using AI to remove the
| extra time-wasting fluff can help. There is no book in
| the public library or Amazon that will show the tricky
| steps to repair a Samsung dryer. It's Youtube videos
| uploaded by random people that has information like that.
|
| My friend learned how to use her new Apple Watch by
| watching Youtube videos. Yes, there's official Apple
| documentation but it's hard to learn from a wall of text
| with static screen shots. In contrast, watching youtubers
| do _live demonstrations_ with their fingers manipulating
| the screen accompanied by voiceover narration is easier
| to understand.
|
| EDIT reply to: _> When you've saved "years of watching"
| by using sponsorblock and 2x speed, that's not how you
| use YT. Sponsorblock saves 25% at most._
|
| Your math is incomplete in your characterization of the
| gp you replied to because he also mentioned _" DeArrow
| (de-clickbaits thumbnails and titles)_. That tool saves
| _100% of the watch time_ when the re-titled videos
| instantly informs him he can totally skip it.
| _Algernon_ wrote:
| When you've saved "years of watching" by using
| sponsorblock and 2x speed, that's not how you use YT.
| Sponsorblock saves 25% at most.
|
| To be clear, I am not saying that youtube has no use, but
| the commenter I originally responded to is clearly in
| excess of whatever that reasonable level is.
| satvikpendem wrote:
| How exactly does one "use YouTube," in your opinion then?
| You also forgot the effects of the other extensions than
| just SponsorBlock, even 2x speed (I watch at higher than
| that usually) saves 50% immediately. You presume I don't
| also do those things you listed, I now have _more_ time
| to do them because of the time saved, not less.
| natebc wrote:
| You watch and enjoy the content that you like, find
| useful or otherwise enjoy instead of min-maxing your way
| to not supporting the platform or creators that are
| spending their time creating that content.
| satvikpendem wrote:
| If the platform and creators are incentivized to waste my
| time, then why would I support that? I watch and enjoy
| the content even more knowing that I am not wasting my
| time.
| natebc wrote:
| This is where the snide touch grass comment above was
| awkwardly pointing. If you feel like you're wasting your
| time and can't be bothered to do so in a holistically
| responsible way (supporting the creators at the very
| least) then perhaps your time would be better spent doing
| something else.
|
| i.e. you may be wasting your time already.
|
| I probably shouldn't have wandered into a thread that's
| ostensibly about ripping off content creators to begin
| with. I'll take my moral high horse elsewhere lol.
| satvikpendem wrote:
| Why would I care about responsibility when I waste my
| time, much less "holistic responsibility?" Sounds like
| your moral systems are different from others' in this
| thread.
| natebc wrote:
| I totally agree, you're absolutely right about that. have
| a good one.
| gosub100 wrote:
| They shared a list of plugins that might be useful to
| other people, so the comment has value. I recently found
| out about the "press to 2X speed" and I love it. Does
| that mean I should go touch grass because I use it in a
| non-stock way? Or are you the ultimate decider on the
| maximum limit of plugins or hacks?
| TheSpiceIsLife wrote:
| It's the seventh commandment, don't you know?
|
| 7. Thou shalt not use more than two browser plugins.
| meroes wrote:
| Ya we are sick and addicted.
|
| Not only do I need the above 4, I need a separate browser
| because YouTube breaks with Firefox, private tabs to not
| taint my preferences when I watch one-ofs, and other addons
| such as a channel and video blocker.
| rcfox wrote:
| YouTube has a speed changer built in. I'm curious what your
| extension would add.
| powerhugs wrote:
| SponsorBlock specifically blocks sponsor segments in YT
| videos.
| wazoox wrote:
| Not the op but the extension I'm using adds 3 very mportant
| features :
|
| * default to my preferred speed (2x)
|
| * keyboard shortcuts to change speed on the fly as needed
|
| * arbitrary speeds up to 4x
| fransje26 wrote:
| > keyboard shortcuts to change speed on the fly as needed
|
| Speed up: >
|
| Slow down: <
| wazoox wrote:
| Undiscoverable, and not configurable.
| eMPee584 wrote:
| with youtube enhancer (there's also an open source one i
| wanted to check out but haven't yet) f.e. one can add a
| speed icon in interface on which the scroll wheel works,
| allowing dynamic speed adjuistment which, accompanied with
| [cursor left] to jump back a bit and [number keys] to jump
| to 0-90% position is amazing for extracting info. Also,
| muting videos, turning on captions and rushing through them
| without the mindfsck is a soul saver often times. Our
| current attention economy with its commercialization
| incentives hopefully is just a toxic glitch in history..
| satvikpendem wrote:
| Higher speeds than 2x. I usually watch at 3x and sometimes
| more (sometimes with captions turned on). Keyboard
| shortcuts as well, as others have mentioned.
| arccy wrote:
| the built in one has keyboard shortcuts too: < and >
| JoeDaDude wrote:
| RIP the Ad Accelerator, which would up the speed upon
| detection of an ad. Youtube has found a way around it and
| it no longer works.
|
| https://github.com/rkk3/ad-accelerator
| GrumpyNl wrote:
| Or ask the original creator if he can publish the script used.
| xandrius wrote:
| You described SponsorBlock, really a game changer. Also works
| on mobile!
| kragen wrote:
| Wow, I'm really surprised that my comment describing 95% of
| YouTubers as "some dumbshit" got voted up to +19. I guess I'm
| not the only old man shaking his fist at the surveillance
| capitalism incompetent confident shouty bullshitter cloud?
| czk wrote:
| Got a chuckle from that recent commit message. I feel you
| ofcrpls wrote:
| The commit in question
| https://github.com/stong/tldw/commit/0144a4e930425d308704de0...
| henry2023 wrote:
| What if we summarize all the information in the world into a few
| hundred volumes of human knowledge, then summarize those into a
| 10,000 pages book, then that into a 10 long form essays, then
| those into a 100,000 chars blog post, then that into a pamphlet
| and finally we summarize one more time into a single tweet.
| userbinator wrote:
| The answer will be a single number, 42.
| henry2023 wrote:
| Then summarize it one last time into a single bit. I like to
| think it'd be '1'.
| gloosx wrote:
| Then this super-condensed 1 explodes and there is no
| information left, just noise
| TeaVMFan wrote:
| 1 in base 42.
| xlii wrote:
| ...but who will have the question?
| PaulRobinson wrote:
| I know you're making a joke, but more seriously I think most yt
| videos have atrocious signal/noise ratios so information
| compression is likely very useful. Less so for many academic
| papers (although they have some pretty awful filler sometimes).
| debeloo wrote:
| This is exactly the problem. There are so many 20 minute
| videos that should have been 2 minutes.
|
| In a way, it's much easier to make the 20 minute video. Just
| hit record, rant an rave, stop recording and publish.
|
| There are indeed justified long videos stuffed full with
| knowledge, insight and witty comments to make it fun.
|
| Then there are "slow" videos but magical. Paul Sellers has a
| 30 min video on how to make mortise and tenons joint with
| hand tools. Just you and him in real time. You get a
| (recorded) private lesson from a master craftsman. It's
| magic. Every minute of it is knowledge transfer.
|
| https://m.youtube.com/watch?v=aBodzmUGtdw
| crakhamster01 wrote:
| Some people inflate their video durations intentionally, but
| I think the majority of people truly think they're using the
| time wisely. Have you ever tried making a quick travel vlog
| of a vacation and ended up with a 15 minute short film? That
| B-roll at the airport was definitely critical to include!
|
| I think the reality is that there are a lot of amateur video
| creators. Elevating the few talented creators through social
| engagement metrics isn't perfect, but I think it works well
| enough. Or at least more so than what these anodyne
| summarizations would give us.
| al_borland wrote:
| I was on YouTube a few weeks ago and saw a 20 minute video
| with a title that looked interesting. Under it was an AI
| summary that saved me 20 minutes, and had me skip the video
| completely. I wish that was under every video.
|
| This week I got a notification about the AI added to YouTube
| to allow users to ask questions about a video. I haven't had
| a chance to use it yet, but I can see that also being useful
| to get the main points from a long video. Up until now, I
| mainly use the popularity indicator on the progress bar.
| Since I watch most videos on my TV, it's harder to use the
| AI, as I would need to pull out my phone, open the same
| video, and ask... that's a bad workflow.
|
| I do find it a little ridiculous that we need AI to summarize
| long videos full of fluff, when the only reason they are full
| of that fluff in the first place is YouTube's own
| monetization policies which pushed the average video from 2-4
| minutes to 10 minutes.
| fifilura wrote:
| Some times I think that would actually be useful for some
| politicians that do not care about history and prior knowledge.
|
| * Rule of law is a good idea
|
| * Dictatorship is a bad idea
|
| * Allowing Germany to occpy Sudetenland in the Munich
| appeasement 1938 was a bad idea. [1]
|
| * ...
|
| [1] https://snyder.substack.com/p/appeasement-at-
| munich?triedRed...
|
| But that said! If this service works I think I could use it. I
| can handle long articles, but have no time to watch YouTube
| clips.
| someothherguyy wrote:
| > into a single tweet
|
| It would say something like, "This text attempts to summarize
| the entirety of human knowledge".
|
| Still, IMO summarizing videos is useful. Even if the summary is
| not accurate or a 1:1 representation of the content, you can
| mostly get the gist of what is being said without being baited
| into watching advertisements.
|
| Although, this site doesn't seem to do a great job at
| summaries. Kagi's universal summarizer has much better results,
| https://kagi.com/summarizer/index.html . However, it requires
| transcripts to be available for videos.
| Sakos wrote:
| I think a lot of people are sort of missing the benefit of
| something like this.
|
| How do you read a book effectively? You skim the table of
| contents. You skim the contents of each chapter and mark
| interesting paragraphs. Then you go through the book another
| 1-2 times, each time getting deeper into the text and cross-
| referencing information between different parts of the book.
|
| What tools like this will do is allow us to apply this same
| workflow to videos, which can greatly enhance our
| understanding of videos we're interested in and help us
| contextualise it with the rest of our knowledge.
|
| I've already been doing this and it's helped me expand my
| knowledge and understanding in ways that wouldn't have been
| possible without an unreasonable investment of time and
| effort.
| 0x073 wrote:
| Like "the book"?
| flemhans wrote:
| Not a single tweet, but 10 brief sentences, as per the AI
| overlords:
|
| 1. The universe is vast, mostly empty, and runs on fundamental
| laws that we barely understand but exploit well.
|
| 2. Life is a self-replicating, entropy-defying phenomenon that
| emerged through chemistry, evolved through selection, and
| adapts through intelligence.
|
| 3. Humans are social primates who dominate the planet through
| cooperation, tool-making, storytelling, and an insatiable drive
| for meaning.
|
| 4. Societies form through shared beliefs, laws, and trade, but
| oscillate between progress and collapse due to power, greed,
| and ignorance.
|
| 5. Technology is humanity's amplifier, accelerating knowledge,
| comfort, and destruction in equal measure, with unintended
| consequences at every turn.
|
| 6. Economies are trust-based systems of resource distribution,
| prone to cycles of boom, bust, innovation, and inequality.
|
| 7. Morality is a human construct, evolving with culture, often
| conflicting between collective well-being and individual
| freedom.
|
| 8. Knowledge is a fractal--deeper the dive, more there is to
| know--yet most wisdom is rediscovery of old truths in new
| contexts.
|
| 9. The future is uncertain but shaped by the tension between
| human ingenuity and our own worst tendencies.
|
| 10. The meaning of life? Whatever gets you up in the morning
| and lets you sleep at night.
| silvestrov wrote:
| > entropy-defying phenomenon
|
| _entropy-exploiting_ phenomenon
|
| is a much better description as life does not defy any
| fundamental laws.
| qiine wrote:
| nice lifr tldr
| jowea wrote:
| https://disney.fandom.com/wiki/Junior_Woodchucks_Guidebook
| rayalez wrote:
| Tried asking Claude to do that, ended up with something pretty
| beautiful:
|
| Everything is made of atoms & energy, life evolves, math
| describes reality, knowledge builds on itself, humans need each
| other & Earth to survive - test ideas, learn from mistakes, be
| kind, stay curious.
| sethammons wrote:
| 42
| guybedo wrote:
| Gemini's output:
|
| Our understanding of reality is fundamentally shaped by the
| power of stories and narratives.
|
| Humanity constantly seeks to impose order and structure on the
| world through systems and frameworks.
|
| The inherent human drive to create and innovate defines our
| art, technology, and design.
|
| We are bound by the complex interplay of connection, conflict,
| and cooperation in our relationships.
|
| Time's relentless flow drives change, progress, and the
| unfolding narrative of history.
|
| The vastness of the unknown perpetually challenges and defines
| the limits of human knowledge.
|
| The search for purpose, values, and meaning is a central and
| ongoing human endeavor.
|
| Abstract concepts and models are powerful tools for
| understanding and navigating reality.
|
| All living things are interconnected within a complex web of
| life and ecological relationships.
|
| The future of humanity presents both boundless potential and
| significant challenges to overcome.
| hollerith wrote:
| That's really bad, but also excellent in a particular way,
| which we might call glibness.
| prvc wrote:
| Getting a "Too long video" message in response to a query is
| frustrating to the user, and, well, redundant information, so
| contrary to the purpose of increasing information density.
| tikotus wrote:
| Something that could be interesting is processing all videos in
| some subgenre, like say "game development", rank them by
| popularity, summarize them, and then analyze for patterns. This
| would be valuable information for anyone wanting to make a video
| with higher chances of being shown to people.
| gloflo wrote:
| Or one could not do that and focus on quality instead.
| tikotus wrote:
| That was actually my initial thought for what I would test
| with a tool like this. How much does quality matter? Quality
| of course has many definitions, but in the light of this
| tool, the measurable quality could be "how much novel
| information does this video have?" I'd hope a trend that pops
| out is that content with more information around a focused
| topic would perform better. But as it is, I only have my
| personal preferences and my suggestions list (based on my
| personal preferences) to go by. I'm not a content creator
| myself, but if I was, I might find this kind of analysis
| interesting enough to pay for.
| ddoice wrote:
| Cool! I had thought about doing the same but in a Chrome
| extension that would use Llama 3.2 with WebGPU, but when I tested
| it, it was very slow and sometimes crashed the browser.
| gloosx wrote:
| Idea hackneyed since LLM's appeared. Cool that implementation is
| open-source, though yt automatic captions are sometimes
| completely off-point, especially when people talking in the video
| don't have a diction of a tv show host.
|
| I wonder if an idea found it's niche after all? Do you guys
| summarise you videos to short texts and that leaves you
| satisfied? For me video is video, I can relax, sit and
| watch/listen to it. With text it is different, it is a mental
| exercise to read and process it, so turning video into text feels
| like an essential downgrade. I would prefer watching at 1.5/2x
| speed instead of text summary if I want to finish it faster.
| nottorp wrote:
| 2x video speed is still a lot slower than text.
| amenhotep wrote:
| If I want to watch a video for the sake of watching a video,
| then no, a text summary would not at all be the same thing.
|
| But most of the time I don't want to watch a video, I just want
| to get information. A text summary then would be strictly
| superior.
| downsplat wrote:
| > For me video is video, I can relax, sit and watch/listen to
| it. With text it is different, it is a mental exercise to read
| and process it, so turning video into text feels like an
| essential downgrade.
|
| Exact opposite for me. Reading goes at my pace, in silence.
| Video is much more invasive, so I avoid it except for the
| highest quality stuff.
| gloosx wrote:
| Interesting, I was always thinking that audio/visual
| information is naturally much easier to consume. For
| instance: I can watch a video and count to 10 in my head at
| the same time - I will still get everything what was in that
| video - but with text it's a much harder task since the head
| is fully occupied with "narrating" the text what I'm reading,
| so reading in the end turns into podcast inside the head
| before actually get consumed.
| throwaway28531 wrote:
| I find it much easier to skim texts to find the information
| I'm looking for. I can't do that with videos unless they
| have clearly marked chapters.
|
| When I read about a new technology and they have a video
| instead of a blog post, I just close the browser. If it is
| useful, someone will probably post a blog on HN at some
| point in time
| righthand wrote:
| > Idea hackneyed since LLM's appeared.
|
| It's an idea that's been around long before LLMs. Check out
| Yahoo under Marissa Mayer acquiring a news summary app. Though
| it is still hackneyed.
|
| https://finance.yahoo.com/news/yahoo-acquires-summly-app-150...
| m1keil wrote:
| That's pretty useful. I think yt itself is experimenting with
| similar feature as I see some videos have AI summaries in the
| official app.
| andai wrote:
| Yeah. "Some." I've seen _one_ in the past two months.
| nottorp wrote:
| Most videos would do well with a summary of the summary of the
| generated transcript. Should take 30 seconds to go through.
| someothherguyy wrote:
| I would like to see one of these that produces an outlined
| summary then has timestamps as citations for the outline's
| individual entries.
| Timwi wrote:
| That would be brilliant. I notice that in math/programming
| videos it will often just say "using a mathematical
| trick/concept" or "using bit manipulation" when it could just
| include the actual formula/equation. It would be nice to be
| able to go to the relevant section in the video to find it.
| andai wrote:
| Change the prompt to "in bullet points" to increase the value by
| about 80%.
|
| (And add "by section" to get to 100%!)
| crocowhile wrote:
| This is so fun!
|
| Could you make it compatible with openrouter please? I see you
| use the openai binding that allows you to specify a different
| base_url so that would work out of the box and integrates well
| with the .env
|
| ``` from openai import OpenAI client = OpenAI(
| base_url="https://openrouter.ai/api/v1",
| api_key="<OPENROUTER_API_KEY>", ) ```
| Brajeshwar wrote:
| The sweet spot of most YouTube videos was supposedly under 10
| minutes. What happened? These days, the typical YouTube videos
| are almost 30 minutes long. I'm fine with documentaries and the
| like. Is there a way to bring back the idea that under 10-minute
| videos work the best?
| bckr wrote:
| You would have to convince YouTube that 10 minute videos
| generate more revenue.
|
| They've been tuning this beast for decades now.
| _Algernon_ wrote:
| It's just a natural evolution of more people watching on their
| TVs and Mobile. Long slop and short slop with nothing in
| between. Must adapt to the TV and phone brain rot zombies.
| ghaff wrote:
| Non-fiction text isn't that different. You have 5 to 10
| minute blog/article reads and 250 page books from publishers
| with not a lot in between. (Yes, you can and I have, self-
| published books of an intermediate length but publishers
| don't generally want anything under 250 pages or so.)
| _Algernon_ wrote:
| You're reading the wrong non-fiction books then. Find books
| that are closer to academia than whatever you find in the
| non-fiction shelf of an airport bookstore.
| geerlingguy wrote:
| 8 minutes is the cutoff for "mid-roll" ads (you can insert ad
| breaks into the middle of your content, not just the beginning
| and end).
|
| But people also realized, especially in the past couple years,
| YouTube seemed to tweak the algorithm to favor watch time more
| strongly than in the past, so people started "fluffing up"
| videos to 20, 30+ minutes. As long as they could get a click
| with a decent thumbnail and title, and enough people to watch
| for at least a few minutes, the average watch time would go up,
| which was a very positive signal. Plus, they could insert like
| 3+ ad breaks (some big creators insert 6+!) and make more
| money.
|
| However, in the past year it seems like YouTube's tweaked the
| algorithm again to not favor watch time as strongly over some
| other metrics, leading to shorter videos being able to get more
| views sometimes (and not just music videos). (But they're still
| "less monetizable".
|
| I think this is why podcasting (and especially extremely long
| podcasts) have gotten so popular. A few thousand viewers
| watching for 2 hours gives a lot more ad revenue (especially if
| you fill in that space with read ads) than like 1 million views
| on a 3 minute clip.
| thdhhghgbhy wrote:
| Text summarisation already frequently misses things, sometimes
| gets them completely backwards. Can't convince myself this will
| be any better.
| TNWin wrote:
| I tried it to summarize some lecture videos. And the summary
| ranged from average to bad. Nothing I couldn't already get from
| the description. Even ChatGPT 4o spits out far better content.
|
| So far my method has been to take the transcript and use an LLM
| with customized prompt for summary.
| computerthings wrote:
| I tried it on music videos, and with most of them I got the
| distinct impression it's drawing from the description or external
| info, maybe PR, because it was very generic, the funniest I saw
| being https://tldw.tube/?v=kxstMJY2Q28
|
| I was about to give up because after all, this isn't for music
| videos, tried something it could not _possibly_ do well with, and
| I 'm impressed, against my will.
|
| https://tldw.tube/?v=7X62OJG6Whg
|
| This isn't just an acceptable summary, this is really good,
| considering how much could be missed (like it did with
| https://tldw.tube/?v=6k7kk1lUnis for example).
| joshuaturner wrote:
| From some light testing, the results here are pretty lackluster.
| I've used videosummarizerai.com in the past via ChatGPT and the
| ability to specify the format of the information you're looking
| for is a huge plus.
|
| Here's the summary of this video
| [https://www.youtube.com/watch?v=W1WkG8WuRcg (AoE2 video)] from
| tldw.tube:
|
| > The video discusses the most powerful team combinations in the
| "10x Shared Civ Bonus" mod for Age of Empires II. It highlights
| how certain civilizations, when combined strategically, can
| create overpowering unit types and tactics. The focus is on four-
| player team compositions that leverage unique benefits: for
| instance, the combination of Franks or Incas for free castles,
| Japanese and Tatars for rapid trebuchet fire, and Celts for
| enhanced onagers. The video also explores various combinations
| for cavalry and infantry, including the Goths for swift unit
| production and a range of secondary civilizations to amplify
| strengths. Commenters contributed ideas for clever combos,
| emphasizing the limitless potential for exploiting the mod's
| mechanics to develop broken strategies that can quickly overwhelm
| opponents by accelerating unit production or enhancing existing
| attributes, thereby showcasing the humorous and chaotic nature of
| the mod's gameplay.
|
| and here's the summary of the same video from
| videosummarizerai.com via ChatGPT with the prompt "Give a summary
| of all civ combos mentioned in order":
|
| > Here's a summary of the civilization combinations mentioned in
| the video in order: Castle Spam Team:
| Incas: Free castles and massive villager armor bonuses.
| Teutons: Adds +30 range to castles. Celts: Castles
| fire at four times the usual rate. Sicilians or
| Spanish: Increased castle build speed (castles built in 30
| seconds). Ultimate Trebuchet Team:
| Japanese: Faster fire rate and instant packing/unpacking for
| trebuchets. Tatars: Adds +20 range to trebuchets.
| Britons: Adds splash damage and perfect accuracy with Warwolf.
| Saracens or Celts: Additional attack bonuses or faster firing
| trebuchets. Best Onager Combo: Celts:
| Increased fire rate and durability for onagers.
| Mongols: Super-speed movement for onagers. Koreans:
| Adds +10 range and removes minimum range. Slavs:
| Makes onagers free. Scorpion Team:
| Khmer: Adds range and additional projectiles. Romans:
| Reduces cost and increases attack rate. Ethiopians:
| Adds torsion engine tech for better area damage.
| Chinese: Adds huge damage to projectiles for almost quadrupled
| effectiveness. [goes on to list remaining 4 team
| combos]
| doright wrote:
| Seems to use gpt-4o under the hood. I wonder if using Deepseek
| would make any difference in quality.
|
| I used to have nice and detailed summaries with this app lasting
| many paragraphs. It used to be totally free and you could submit
| as many links as you wanted. However they started forcing you to
| wait dozens of hours between summaries or pay for credits. I
| haven't found a YouTube summarizer as high quality since.
|
| https://play.google.com/store/apps/details?id=com.emote.yout...
|
| I'm thinking one could replicate such a service pretty easily and
| be able to plug in your Deepseek API token instead. It would be
| convenient if such services let you "bring your own API key" so
| to speak, so you'd only have to manage one bank of credits. But
| it's understandable that lots of people want in on a slice of the
| AI pie right now.
| ruibiks wrote:
| Hey, try my app https://Cofyt.app to see if you like it and
| leave your feedback here in thread for others to read.
|
| It's excellent for summaries but also answers your questions
| grounded in the video.
|
| It's free!
| jisnsm wrote:
| I tried using it on a 2-hour video and it said that... the video
| is too long?
|
| Kind of like hiring a painter to paint your house and him
| refusing the job because your house is unpainted.
| haykuro wrote:
| great analogy.
| CM30 wrote:
| Tried it on 3 random videos I watched, and the results were...
| mostly good, albeit mixed.
|
| On the one hand, it got my video about a Mario & Luigi:
| Brothership glitch dead right, immediately listing where you'd
| need to die to get an item early and what you'd get out of it.
|
| It also did an okay job summarising a Zelda dungeon analysis
| video by someone I'm subscribed to, with some info on why that
| dungeon was well-designed that clearly came from the video.
|
| Unfortunately, it did a poor job at summarising a video about
| plagiarism in the YouTube speedrunning essay space, associating
| the problem with smaller creators rather than the person the
| video was about and leaving out far too many details to be
| useful.
|
| This seems to confirm my assumptions about how an AI summariser
| would work in general; if the original media is a straightforward
| piece about one easily understandable topic, then it'll do fine
| and work about as well as a human would. If it's a longer piece
| with multiple points backed by various examples, then it'll
| struggle to summarise it in a way that makes sense.
| braiamp wrote:
| > If it's a longer piece with multiple points backed by various
| examples, then it'll struggle to summarise it in a way that
| makes sense
|
| I've found the same problem with humans too, so it's not like
| an improvement over humans.
| atommclain wrote:
| Curious how well this compares to the same functionality built
| into Gemini.
| Awtem wrote:
| I like the idea, but I get the message: "Too long video"... So I
| am wondering about its application if it only works for short
| videos
| hrkname wrote:
| I use YouTube's transcript feature or 2x speed for slow talkers.
|
| Especially with political commentary many YouTubers stretch their
| content or talk slowly for more ad deliveries. It is annoying,
| because often there is an important five minute section hidden in
| a one hour video.
| tanewishly wrote:
| Interestingly, YouTube also lends itself for this sort of thing
| for movies. You can watch scenes from movies and get the gist,
| while only spending ~30min on a movie. It's a great way to watch
| mediocre movies - they're not so horrible (1), and you do get the
| entertainment value without being exposed to the shlocky-est
| parts.
|
| (1) Eg, if you skip the beach scene, Terminator Dark Fate is...
| palatable! But yeah, the reasons for mediocrity typically still
| shine through a bit.
| nashashmi wrote:
| What is really cool about this is that my attention span is kept
| after reading the summary.
|
| Like the following video:
| https://m.youtube.com/watch?v=JBkAKCJJuFQ
|
| I am not getting emotionally invested to the point of losing the
| end focus of the speaker. Actually now I am looking forward to
| it.
| dankobgd wrote:
| too long even this site cannot "watch" it
| wopwops wrote:
| "Too long video" lol
| esafak wrote:
| Does it just parse the captions like every other service or does
| it analyze the audio too?
| leovander wrote:
| > An error occurred: tuple indices must be integers or slices,
| not str
| stong1 wrote:
| This should be fixed!
| albert_e wrote:
| Tried Andrej Karpathy's latest video --
| https://www.youtube.com/watch?v=7xTGNNLPyMI
|
| It said "too long video"
|
| Exactly, duh ;D
| acrophiliac wrote:
| An error occurred: tuple indices must be integers or slices, not
| str
|
| https://www.youtube.com/watch?v=Ks-_Mh1QhMc&list=PLWDUzz3hCD...
| stong1 wrote:
| This should be fixed now. There was temporary outage due to
| proxy running out of bandwidth.
| stong1 wrote:
| Hi HN! I'm the author of this service. Thank you for your
| support.
|
| There may have been some temporary downtime due to residential
| proxy running out of bandwidth. I have purchased additional
| bandwidth. (I run this service for free.)
|
| There also may be some errors with particular videos because they
| are not accessible in certain regions. For now all requests to
| YouTube originate from United States, but open to change in the
| future to some kind of round-robin or fallback system.
|
| I know it's not perfect. I developed the tool originally for my
| own use. It's open source and I'm open to any patches or pull
| requests.
|
| Enjoy!
| JTyQZSnP3cQGa8B wrote:
| > open-source
|
| > OPENAI_API_KEY
|
| Choose one.
| SomeoneOnTheWeb wrote:
| The service itself is open source ; but it relies on a closed
| source service. Both are not incompatible.
| suprjami wrote:
| > OPENAI_BASE_URL
|
| > OPENAI_API_HOST
| NewUser76312 wrote:
| Hey this is really cool, I literally had the same idea about a
| month ago but ultimately decided to not pursue it. Glad someone
| else did.
|
| A few quick Qs -
|
| 1) Do you use the available auto-generated transcripts from
| youtube? Or do you do any audio parsing? I know transcripts
| aren't always available.
|
| 2) Do you have any plans to monetize in some way, do you think
| it would be possible? It's definitely a neat product but a tad
| generic and replicable, so I'm curious.
| arccy wrote:
| how expensive is the bandwidth if you're buying from
| residential proxies?
| ANighRaisin wrote:
| How does this compare to gemini thinking experimental with
| apps?
| fullshark wrote:
| So what is everyone doing with all this free time they've now
| accumulated from not reading, watching, or listening to media?
| righthand wrote:
| Chatting with AI bots.
|
| I agree that this mentality of works being "too long so don't
| ingest it", is not a healthy way to go about life and thinking
| in general.
| DennisP wrote:
| It's not that, it's "too long with low information density so
| ingest it more efficiently." That way we can spend our time
| on things that are more productive or enjoyable.
| righthand wrote:
| > That way we can spend our time on things that are more
| productive or enjoyable.
|
| That seems like an arms race, how do you know there aren't
| more productive ways to enjoy your time? Can something less
| productive be more enjoyable?
| DennisP wrote:
| I don't, and yes. I suspect this varies by person.
| BeFlatXIII wrote:
| Sleeping. Occasionally hitting a drum.
| tonyhart7 wrote:
| we automate so much, we end up making human redundant
| renewiltord wrote:
| There's also a funny one like https://tldw.tech/ which makes
| everything into a The Verge video.
| bli940505 wrote:
| I tried a video 1:42 hours long and it told me it was too long.
| Increase the limit?
| JoeDaDude wrote:
| This is not the only service of its kind. There is also the Video
| Gist:
|
| https://www.videogist.co/
|
| I can't speak to the differences between the two. Once casual
| observation is that the one posted by the OP seems to be a lot
| faster..
| ggasp wrote:
| I made a shortcut for iPad to do more or less the same thing and
| promised myself to keep "engineering" the prompt but work pretty
| well as it is.
|
| Link:
| https://www.icloud.com/shortcuts/fbb5a315cb354fbf903bcfa4f40...
|
| What the shortcut does is taking the transcript of the video and
| asking your ChatGPT app to extract main ideas, quotes, and facts.
|
| The prompt comes originally from fabric CLI tool
| https://danielmiessler.com/blog/fabric-origin-story
___________________________________________________________________
(page generated 2025-02-15 23:01 UTC)