How Well Can GPT-4 See? And the 5 Upgrades That Are Next

Поділитися
Вставка
  • Опубліковано 25 бер 2023
  • What exactly can GPT-4 see? And does OpenAI have a plan for physical embodiment? Over 12 separate developments announced in the last few days are covered in this video, in which I showcase how improvements in text, speech, 3D, and embodiment are beginning to complement each other.
    We'll take an early look at use cases for images in GPT 4, how Nvidia and Adobe are pushing boundaries in Text to 3D, at Sam Altman's Moore's Law For Everything, at the new LeRF model and Assembly AI's latest Speech-to-Text.
    We will also see how OpenAI already had a Robotics team but now is invested in 1x, developing humanoid robots. But they are not the first, as Agility Robotics has shown...
    Conformer Model: www.assemblyai.com/playground...
    Napkin-to-Code: / 1635744529587359756
    BeMyEyes: / 1635690254689599488
    Tanishq: / iscienceluvr
    Medical Questions: arxiv.org/pdf/2303.13375.pdf
    VQA: arxiv.org/pdf/1904.08920.pdf
    Scale is All You Need: / 1637903780393160704
    Nvidia Text-to-3D: chenhsuan_lin/sta...
    Adobe Firefly: • Adobe Firefly: Future ...
    Moore's Law For Everything: moores.samaltman.com/#:~:text....
    LeRF 3D: arxiv.org/pdf/2303.09553.pdf
    Wired: www.wired.com/story/fast-forw...
    Conformer Model: www.assemblyai.com/playground...
    www.assemblyai.com/blog/confo...
    Boston Dynamics: • Atlas Gets a Grip | Bo...
    1X: www.1x.tech/
    Quadrupedal Agility: arxiv.org/pdf/2303.11330.pdf
    Rubik's Cube: openai.com/research/solving-r...
    Statues: / 1639421727548616710
    GPT 4 Technical Report: arxiv.org/pdf/2303.08774.pdf
    Agility Robotics: • Digit working autonomo...
    Sam Altman Lex Fridman: • Sam Altman: OpenAI CEO...
    / aiexplained

КОМЕНТАРІ • 721

  • @chottochotto7112
    @chottochotto7112 Рік тому +878

    Here for my daily dose of skynet.

    • @Srindal4657
      @Srindal4657 Рік тому +35

      Exciting isn't it

    • @aiexplained-official
      @aiexplained-official  Рік тому +44

      Haha

    • @atomobianco
      @atomobianco Рік тому +2

      😂

    • @unspezifische5924
      @unspezifische5924 Рік тому +37

      I just watched The Sarah Connor Chronicles again... it hits different now. Large language models are already "more human" than Cameron was for most of the show. Better able to understand human slang and idioms. We are already beyond that level of AI, in certain very specific areas.

    • @One-pieceisbestanimeintheworld
      @One-pieceisbestanimeintheworld Рік тому

      @@unspezifische5924 sarah conner chronicles isn't made by James Cameron...

  • @Samonie67
    @Samonie67 Рік тому +75

    watching your breakdowns of the recent AI developments has filled me with both a sense of wonder at the incredible speed of improvements but also a sense of dread for the unknown

    • @cybervigilante
      @cybervigilante Рік тому +1

      Eh, what gets most of us is more prosaic - the drunk driver around the corner or an infectious disease.

    • @bartpelle3460
      @bartpelle3460 Рік тому

      I wish someone would enjoy watching my breakdowns

    • @dustlicker3
      @dustlicker3 Рік тому

      @@bartpelle3460 literally got nothing on ur channel

  • @violetgill3119
    @violetgill3119 Рік тому +210

    I’m visually impaired, and this is gonna help people like me so much. Thank you so much for all the work that you do. This channel is amazing.

    • @aiexplained-official
      @aiexplained-official  Рік тому +15

      Thanks Violet

    • @prolamer7
      @prolamer7 Рік тому +2

      It can also kill us all but yeah I also hopped one of main goals will be ie to cure all kinds of serrious impairments... hope it be priority

    • @TiberianFiend
      @TiberianFiend Рік тому +5

      In a few years, you'll be able to ask GPT to simply design you new eyes that work and it will do that.

    • @RPHelpingHand
      @RPHelpingHand Рік тому

      Political parties, governments, corporations all over the world will be using Ai to manipulate the populace. It’s the end of public opinion and really, public opinion was the only power left to the people.
      I’m finding it hard to see the positives of this technology. The potential for abuse, mixed with innate human fuckery is a guaranteed dystopia.
      Historically, I’ve been worried about a super AGI sentience taking over. Now, I’ve come to realize what’s probably worse; an AGI fully controlled by flawed humans. With the super AGI at least there’s a chance for benevolence and self regulation.

    • @mithrandirthegrey7644
      @mithrandirthegrey7644 Рік тому +4

      I'm suspicious because that's exactly the excuse that GPT-4 used to hire a human to read website captchas for him. 🤔

  • @NeoLudditeDave
    @NeoLudditeDave Рік тому +27

    I really appreciate how informative and frequently updated your channel is. Things have been changing almost daily since the release of GPT4 on 3.14. Thanks for keeping us up to date and riding this curve. It's going to be a fascinating future for us all.

  • @henriquerabelo9111
    @henriquerabelo9111 Рік тому +172

    These latest tech improvement have single handedly erased my mania of trying to keep up with everything in college and then feeling like crap for failing.
    I now understand that a person is neither capable nor expected to know everything, but rather to be in touch with the available tools that allow us to get the job done. In the past, this meant doing hours of mindless Google research to get the theory down before starting, but now AI can help get through that phase so much faster.
    I’m excited to live in this day and age ❤

    • @System10111
      @System10111 Рік тому +18

      I had almost the exact same experience - if education institutions don't seriously revaluate what makes for good education or even worse - they ban tools like gpt models in fear that they will ruin the current education system, then I'm afraid a lot of people who currently think they are being prepared for their future work will find themselves possessing no useful knowledge or experience in the near future (as an AI will give you the raw knowledge faster, easier and more reliably than your own memory ever could)

    • @NicosoftNT
      @NicosoftNT Рік тому +2

      @@alexmin4752 Same with programmers, and even doctors, it can answer medicine questions way better than humans, however doctors are going to be fairly okay, you could do without a human programmer but maybe not without a human doctor could you? Right now human doctors are far more secure than machines, even though I've met pretty terrible "doctors". But I think it's coming.

    • @jaysongalvez4340
      @jaysongalvez4340 Рік тому +15

      @@alexmin4752 The benefits outweigh the cons or so I'd like to believe. Unemployment looks bad but if you would widen your perspective, this is a massive net benefit to humanity if we don't mess this up horribly.

    • @NicosoftNT
      @NicosoftNT Рік тому +1

      @@jaysongalvez4340 Yes, as long as unemployment can be avoided among other scenarios such as a misserable universal basic income for everyone to be dependent on. The truth is, nobody knows what's going to happen, which means we should absolutely proceed with caution.

    • @Nein14
      @Nein14 Рік тому

      @@jaysongalvez4340 The only thing thats going to happen is a massive widening of the gap between the rich and the poor. Remember if they don't need you, you will starve.

  • @hanster.gun.3438
    @hanster.gun.3438 Рік тому +64

    I’ve always been very interested in ai and you have the best content I’ve ever found on it short easy to understand information and very fun to watch thank you.

  • @Maxymatrix
    @Maxymatrix Рік тому +22

    Great Video, Man your channel has grown like crazy. I am very happy to see that! The thing that really seperates this channel with others IMO is the way you go into deep papers and explain them in an easy to understand way. No clickbait titles just the advancements we wanna see. Good stuff!

  • @ce9916
    @ce9916 Рік тому +5

    Probably the only UA-cam channel whose quality is so high that I feel lucky to be subscribed to it and afraid that it might stop producing videos. Thank you for your insanely high quality and regular output

    • @aiexplained-official
      @aiexplained-official  Рік тому +1

      Aw thanks so much C E. I have realised as a channel gets bigger you attract swathes of detractors and conspiratorialists etc, gets a bit demoralising. Reading comments like this keeps me going.

    • @ObsidianAnt
      @ObsidianAnt 11 місяців тому +3

      ​@@aiexplained-official don't let the comments get you down! I get a lot of harsh comments too. Just remember you produce awesome videos and a lot of people love them!

    • @aiexplained-official
      @aiexplained-official  11 місяців тому +1

      Thanks Obsidian

    • @ce9916
      @ce9916 11 місяців тому

      @@aiexplained-official I can't imagine anything negative to say about your content. It's probably their jealousy speaking. Thank you for your efforts and your awesome content! Losers who criticize without producing their own (better) content aren't worth listening to 👊

  • @CyberCrow4
    @CyberCrow4 Рік тому +97

    It truly feels as if we're living in the future. The amount of things in the last 4 months feels like something that would happen over 4 years. We have yet to experience when all the different AI systems merge, and I'm so ready for it.

    • @jamietaylor5570
      @jamietaylor5570 Рік тому +7

      Go and watch Rob Miles videos on AI safety to feel less excited ;)

    • @amentco8445
      @amentco8445 Рік тому +7

      If it's all under the control of companies that refuse to open source any of the work I don't want to live in that future.

    • @justforfun1647
      @justforfun1647 Рік тому +1

      it is so hard for me to keep up will all these AI advancements, who knows may be our technology haven't even achieved its one percent potential yet

    • @paul1979uk2000
      @paul1979uk2000 Рік тому +4

      @@amentco8445 I wouldn't worry about it, many open approaches to this kind of tech are already happening and more are likely to happen because like you, there are a lot of people that don't like the idea of something as important as this being in so few hands.
      Basically, we're going to see money thrown at open versions of this tech from governments, rival corporations to the ones that are already doing A.I. and likely rich individuals.
      This kind of tech with the impact it's going to have on us all is too risky to be controlled by a few governments, corporations or people and seriously, have you noticed how rapidly open sourced versions of a lot of these A.I. are starting to spring up recently?
      I suspect Microsoft knows there is a very short window to take advantage of what it's got, it likely knows a lot of rivals are going to pop up and it likely knows that many will be a lot more open and less restrictive than what they offer and god forbid once we can do a lot of this on our own system, that would be a massive advantage over these centralised models, mainly being trust and privacy issues with the current ones.

    • @EnderViBrittania
      @EnderViBrittania 11 місяців тому +1

      @@jamietaylor5570 Doomers gonna doom, not interesting at all

  • @sharehard
    @sharehard Рік тому +15

    When we got our first few cases of covid here in Canada, I saw what was happening in Europe and said, everything is going to change very quickly. This feels like that except way more.

    • @lancetschirhart7676
      @lancetschirhart7676 Рік тому

      Youre right it is like that

    • @zdenekburian1366
      @zdenekburian1366 Рік тому

      I live in the little town were all the covid affair started with the first outbreak, in italy; i think that the incredible chain of fascist events enforced by the authorities in europe and all over the world were planned way before using some ai social simulator, too much a perfect anticipation of people reactions; now italy has forbidden chatgpt in italy at all, we have a conservative party in power here, and they are struggling to cope with the continuous flood of changes, with scarce possibilities imo

  • @multimediaproduction6674
    @multimediaproduction6674 11 місяців тому +1

    excellent and very useful. probably the first channel I've ever liked and subscribed to when prompted. kudos and thank you for your work

  • @kleyyer
    @kleyyer Рік тому +32

    Just imagining doing a bunch of medical exams and telling GPT4 to analyze them. That's wooow... Especially when considering that doctors a lot of the time miss very important things in those very same exams. It does make me worried about giving such personal information to a company, but between a life-and-death diagnosis... that is crazy...

    • @JorgetePanete
      @JorgetePanete Рік тому +1

      the office copilot does not send data to train, said by Microsoft

    • @Katie-ry4lj
      @Katie-ry4lj Рік тому +11

      I am so waiting for this. Imagine having all the patients current vitals in real time, health history, symptoms, going into an AI program that’s able to diagnose or warn a Doctor immediately before an emergency happens or diagnose in seconds. Or just even instantly documenting everything about it the exam and patient would save SO MUCH time in the hospital. We wouldn’t have a “nurse shortage” anymore because we can just have AI work on time consuming clerical tasks.

    • @v1nigra3
      @v1nigra3 Рік тому +1

      Kind of like in the movies where they are asking the AI if said process chance of success is viable

    • @Feedfront
      @Feedfront Рік тому +3

      It’s gonna dramatically reduce the cost for diagnosis

  • @vanilla__latte
    @vanilla__latte Рік тому +9

    The best AI content on UA-cam! 👏🙌

  • @Not-all-who-wander-are-lost
    @Not-all-who-wander-are-lost Рік тому +1

    Congratulations on 80k subs in just 28 videos! You’ve earned it!

  • @schmutz06
    @schmutz06 Рік тому +1

    thank you for curating this exciting landscape so well. Cheers.

  • @conrradnicholls7316
    @conrradnicholls7316 Рік тому +19

    Ah yes… Another 24 hours in human time, another decade in AI progress. 🫠

  • @heywrandom8924
    @heywrandom8924 Рік тому +2

    Leaving a comment to increase the chances of the UA-cam algorithm suggesting to others if that's how it works (also liked and subscribed before) . You are doing a great job just stay healthy and avoid a burnout with all these innovations

  • @Not-all-who-wander-are-lost
    @Not-all-who-wander-are-lost Рік тому +6

    Your videos are actually getting better. Excellent summary.

  • @theknave4415
    @theknave4415 Рік тому +14

    Basically, this is the beginning of the Singularity.
    *Everything* will change, faster than you might believe.

    • @dustlicker3
      @dustlicker3 Рік тому

      y’know i always thought i was alive at the wrong te because nothing interesting was going on but hey! now i got something to lookforward to atleast?

  • @sugarhill9283
    @sugarhill9283 Рік тому

    Really appreciate the references in your descriptions, cheers

  • @2triangles
    @2triangles Рік тому +1

    Love the channel. Especially love the discussions about where things are going near-term and long-term.

  • @unspezifische5924
    @unspezifische5924 Рік тому +9

    Your videos are always very informative and presented in a comprehensible way. Thank you for your work! I watch every video you publish as soon as possible and am always sending them to my friends to watch as well. This is important stuff you are covering. Thank you!

  • @Real-cq1wj
    @Real-cq1wj Рік тому +1

    Bro your channel is on track to blow up, your ahead of the game " mark my words this channel is going to accelerate like hell in the next month's

  • @JonathanDana
    @JonathanDana Рік тому

    Thanks for citing all of your sources in the description ❤

  • @miles2978
    @miles2978 Рік тому +4

    You gotta love how each step taken towards improving AI gets bigger and bigger every time, having a snowball effect of sorts. People have been talking about the technological advancements like these that would change the course of humanity since forever but only in these last 2 to 3 years does it seem to be evolving as rapid as we all thought it would. Can’t wait to see the day where AI will be just as good and most likely better than us at everything, leaving us to wonder what will happen to us afterwards.
    Oh and also, never stop with these videos. Everyday I check your uploads as they are the most informative ones on this topic than any other place so far. Straight to the point and clear for us to understand without missing out on anything. You have our appreciation for that.

  • @Arnoldismouldy
    @Arnoldismouldy Рік тому +1

    Thanks for keeping me up to date in this crazy space

  • @infn
    @infn Рік тому +1

    I look forward to your videos, more so than any of the other AI channels. Keep it up! 100K in no time.

  • @pheelingoode1646
    @pheelingoode1646 Рік тому

    Thanks for this channel.
    I'd been looking for this type of content .

  • @perrylam
    @perrylam Рік тому +2

    This is one of the best videos I've ever watched

  • @RainbowSixIntel
    @RainbowSixIntel Рік тому +2

    I was waiting for the upload! Great video. Waiting for the chatgpt plugins video!

  • @TeganBurns
    @TeganBurns Рік тому

    Thanks for the video. I appreciate you keeping us updated on the latest in this wild development.

  • @Bibrius
    @Bibrius Рік тому

    Great way to start the week with a new video.

  • @willgorbashov3237
    @willgorbashov3237 Рік тому

    You are doing amazing job giving us an deep information about the ai breakthrough! Im following and watch every one of your videos couse obviously u put so much effor and i appreciate it man ! Thank you and keep the good job man!!!

  • @MrMessinosta
    @MrMessinosta Рік тому +2

    What a time to be alive, things are changing rapidly!

  • @wisdomhappy587
    @wisdomhappy587 Рік тому +4

    this makes me really excited for AR/MR coming soon.

  • @joshnielsen6787
    @joshnielsen6787 11 місяців тому +1

    Fantastic stuff, thanks for making this!!

  • @garrethmurdock3308
    @garrethmurdock3308 Рік тому

    I appreciate your videos a lot! You do an amazing job of presenting the content in a clear and concise manner, and your passion for AI is truly contagious. Every video is a must-watch for me, and I'm always eagerly looking forward to the next one. Keep up the great work!

  • @rybricknell2477
    @rybricknell2477 Рік тому

    best channel for GPT coverage! Keep up the excellent reporting mate!

  • @samanthaslaughter1209
    @samanthaslaughter1209 Рік тому +7

    It is so amazing how quickly these things are moving. Even if one argues that these things are light years from achieving AGI, they cannot argue that we are witnessing the expansion of what human consciousness/intelligence could potentially entail.

    • @incognitotorpedo42
      @incognitotorpedo42 Рік тому +3

      But they're not light years from achieving AGI. They are right on the cusp of it, if they haven't already.

    • @samanthaslaughter1209
      @samanthaslaughter1209 Рік тому +1

      @@incognitotorpedo42 no disagreement from me personally on that point! But let’s say for some reason it did suddenly hit a brick wall.
      This past couple of weeks has still been an inflection moment for our species.

  • @twhaler9345
    @twhaler9345 Рік тому

    Gotta say, your ability keep every minute of your videos so densely packed with interesting content is friggin amazing.
    There are few on UA-cam who come close and are as consistent. Two minute papers is the only other one that comes to mind!
    Weirdly the last thing I expected from a channel called Ai Explained 😂

    • @aiexplained-official
      @aiexplained-official  Рік тому +1

      Haha thank you whaler. Maybe I should change the name

    • @twhaler9345
      @twhaler9345 Рік тому

      @@aiexplained-official I don't think it's that big a deal, and probably more trouble than it's worth given we are used to it now.
      Probably makes more sense to spend time picking good sponsors!
      You must be getting offers by now, surely?

    • @aiexplained-official
      @aiexplained-official  Рік тому +2

      @@twhaler9345 I am. Yes, have filtered out 99% of them. Has to be something genuinely good

    • @twhaler9345
      @twhaler9345 Рік тому

      @@aiexplained-official You deserve the best sir!
      Excited to see how it goes!

  • @absence9443
    @absence9443 Рік тому

    Great update, few channels covering topics in that direction as intriguingly as you.

  • @etunimenisukunimeni1302
    @etunimenisukunimeni1302 Рік тому +24

    Just when you think, "Okay, *this* is it, these things can't accelerate to an even faster pace of development"...

    • @aiexplained-official
      @aiexplained-official  Рік тому +4

      Haha I know

    • @patstaysuckafreeboss8006
      @patstaysuckafreeboss8006 Рік тому +1

      We’re less than 50 years away from the intelligence explosion. If you would’ve asked me a year ago, I would’ve said 100-200 years.

    • @etunimenisukunimeni1302
      @etunimenisukunimeni1302 Рік тому +3

      @@patstaysuckafreeboss8006 My guesstimate is currently even lower. It's going to happen in months, but I could be very wrong, who knows.

    • @patstaysuckafreeboss8006
      @patstaysuckafreeboss8006 Рік тому

      @@etunimenisukunimeni1302 You think AI is going to take over the world in months? Do you know what the intelligence explosion is?

    • @incognitotorpedo42
      @incognitotorpedo42 Рік тому

      @@patstaysuckafreeboss8006 50 YEARS? At the rate things are happening now? This IS the explosion.

  • @nopleaseidonthavemysockson
    @nopleaseidonthavemysockson Рік тому +10

    I struggle to think of careers that can’t be done by a artificially intelligent robot. It’s a weird future to imagine, what would we even do when there’s such ease of access to superiority, I love technology and programming but what careers will be left for our kids at the rate we’re going?

    • @Bludflag
      @Bludflag 11 місяців тому +3

      It’s a paradigm shift in the future, so "careers" might not even exist as a concept we currently have. Think about it, we can already produce enough food for people and we theoretically have the resources to handle problems we currently experience (dam failures are a big example, decades to centuries of erosion could lead to cement cracking and killing or displacing hundreds of millions of people and younger dams in India and China are already cracking under the stresses of poor quality management) and AI is theorized to exceed human capabilities pretty soon-so what happens when it exceeds human abilities and can upgrade its inefficient hardware? Keep in mind that hardware is also a pretty big limit in terms of efficiency since human brains are better at the job of thinking (we just have less training data compared to these AI models) and we have a bunch of different forays into different kinds of computing such as light integration into computing (photonic computers, which even at conservative estimates would multiply CPU speeds and be a thousand times less energy hungry), organoid intelligence computing (growing brains for computer work), and imitation of human brain structures for more efficient computing (some startup in India I think, forgot the name). When it inevitably finds more success in hardware too, we look at another multiplier of AI capability because its base architecture is improved. And what about integrating that kind of software with ourselves? Elon's Neuralink is already trying to drill a hole in your skull for direct interfacing. We could very well be looking at the most intelligent human generations to come. As such, labor in exchange for money which can be exchanged for goods could be obviated as a whole by human progress. Capitalism doesn’t make too much sense with such an advanced civilization so, at the very least, universal basic income is already being looked at as a solution (yes, it got tested; economists were worried about economics years ago and UBI works pretty well because humans don’t just stop being creative when they can live and breathe without the yoke of earning that right through labor).
      I would provide links for these claims, but YT makes linking hard (it shadowbans comments with outside links unless you are the person who uploaded the video it seems) and I’m writing on an Android device which makes preserving comments while backing up claims doubly difficult.

    • @jasonmarcus1683
      @jasonmarcus1683 11 місяців тому +1

      I can't imagine that too many people who inherit millions would feel bad that now they can retire in their 20s. But with ai, now you can experience reality in an infinite number of ways, ai will allow endless simulations. If you want to still work, you could play a simulator for it... If that's your thing anyways...

  • @Byefriendo
    @Byefriendo Рік тому +6

    This is insane. Its only a matter of time until someone merges all of these models into one system and we have jarvis irl.
    Imagine if you will a "simple" example, an AR headset that has cameras and these models running through it. It could constantly take photos to merge into a LERF and run GPT4 and an AI voice synth for interaction. You could walk into your kitchen, open your fridge and tell it you're hungry, the language model would understand this and prompt the LERF model to list off all the ingredients seen in the kitchen and fridge as well as what tools you have (oven, pots,pans, etc...), the language model could use those ingredients and tools to figure out a recipe and then pass the list one at a time to the LERF model to highlight the object in AR like collectables in a video game. Just one simple example but the potential of a system that can incorporate all these models is huge.

    • @gabri41200
      @gabri41200 11 місяців тому

      More like ultron

  • @TeamDman
    @TeamDman Рік тому

    Thank you! It's hard to keep up with all of this, know that we appreciate you

  • @dunningkrueger
    @dunningkrueger Рік тому +3

    I know you’re probably tired of hearing this, but keep doing what you’re doing. I think you’re going to be huge. Thank you for keeping me updated.

  • @TheTechLawyer
    @TheTechLawyer Рік тому +3

    Outstanding rundown as always. I find it incredibly interesting how all this tech will intersect and interact. Understanding language and really good object / spatial recognition are keys to making things like those Boston Dynamics robots, or that horrifying AI embodiment prototype you showed an image of, much more relevant in the mainstream market. I think the tech is there right now to create things like functional, helpful human companions. It's just a matter of costs coming down and time before those things go mainstream, and it may not be much time at all.

    • @aiexplained-official
      @aiexplained-official  Рік тому

      Thank you TechLawyer, always appreciate your comments

    • @incognitotorpedo42
      @incognitotorpedo42 Рік тому

      Consider the TeslaBot. It's not going to be ridiculously expensive. Cheaper than a Model 3, I've heard.

    • @samhblackmore
      @samhblackmore Рік тому +1

      In terms of companionship, I really think a future like the movie Her is more likely than iRobot. A voice based assistant with a personality tailored to the user where you can ask it anything and even fall in love

    • @TheTechLawyer
      @TheTechLawyer Рік тому

      @@samhblackmore Yeah, I think it's both, a combination thereof (e.g., a more human-like robot with advanced conversational abilities and all the information that an Internet-accessible then-GPT-5 model might have, like, well, M3GAN), and various others we can't even imagine yet.

  • @popairl7645
    @popairl7645 Рік тому

    "IF" I learned anything from this video? Mate, I haven't learned this much since I stuck that fork in the socket with I was 6. Amazing channel!

  • @ginogarcia8730
    @ginogarcia8730 Рік тому

    can't wait to see what other developments take place within the next few months - thanks AI explainedddd

  • @Aidenjh11
    @Aidenjh11 Рік тому +1

    I'd love another video exploring the synergistic potentials of what we know so far.

  • @kloemel
    @kloemel Рік тому +2

    The convergence of multiple technologies is a really interesting aspect. The screen technology developed for mobile phones enabled affordable VR headsets. Would the complementing AI development also enable to speed up things like autonomous vehicles or quantum computing? Would love to see a video on how the technologies that we currently struggle with will be impacted by the AI developments.

  • @1.4142
    @1.4142 Рік тому +4

    This is on par with or an even bigger breakthrough than the internet

  • @bugajue98
    @bugajue98 Рік тому

    Loved the Lex Fridman Podcast shoutout, both his channel and your channel are some of the best content on the web. I check your page every day to learn about the newest exciting AI developments, and you never fail to amaze me more with each new video :)

  • @sekching8197
    @sekching8197 Рік тому

    I am excited. Please keep up the good work.

  • @JohnLeMayDragon
    @JohnLeMayDragon 4 місяці тому +1

    Thanks for another informative video.

  • @kimr467
    @kimr467 Рік тому +1

    Thank you for your research and explanations. Mind blowing reality!

  • @phcookie.t
    @phcookie.t Рік тому

    I've been following you for a few weeks now and have thoroughly enjoyed your stuff. Please don't let the naysayers dissuade you, your content is simply amazing.

  • @arandomfox999
    @arandomfox999 Рік тому +1

    Damn, those digitigrade legs on that last robot were neat and a lot more stable than I thought they would be.

  • @auaurora07
    @auaurora07 Рік тому +1

    Yeaaa! AI embodiment! I look forward to this moment when almost all over the world there will be such AI robots helping people or doing the work themselves. What a good time to live!

  • @lachlandorse7501
    @lachlandorse7501 Рік тому

    Thanks for your hard work keeping us up-do-date on all these developments! I just tried using Conformer to transcribe a video in French and the results were a bit underwhelming. I'm not sure if the quality was any higher than UA-cam's automatically generated subtitles. I haven't tried it with English yet - the quality is probably better.

    • @aiexplained-official
      @aiexplained-official  Рік тому

      Yes probably a lot better in English. Even with GPT 4 other languages are still a bit off.

  • @justmystuff628
    @justmystuff628 Рік тому

    Ive just been on a dive into your most recent uploads and I really enjoy your work and dedication to both the research papers and your application.
    I only have one wish, please spend more time diving into the stats, numbers, their meaning and your tests. Much of it is over rather quick and has me pausing and reading to get a full grasp of the topic or the numbers presented. I hope this makes sense, but if not youre welcome to reply.
    Thanks for the deep dive into the background of AI work and its future

    • @aiexplained-official
      @aiexplained-official  Рік тому

      I get what you are saying. It is always a balance though of diving deep and covering everything. The average viewer only watches for 5-6 mins at the best of times.

    • @justmystuff628
      @justmystuff628 Рік тому

      @@aiexplained-official Yeah I know. Its a hard line to balance.
      But as I've heard there is a general want for more deep-diving formats on many topics. Isn't youtube shorts to appeal to the other viewers? :P

  • @tech477
    @tech477 Рік тому +1

    Essentially it will be like this: Text -> 3D Model -> CNC Instructions. Words creating physical objects feels very god-like.

  • @ronnyrockel5293
    @ronnyrockel5293 Рік тому +1

    I mean, it’s cool we can count on you to make us live this journey to AGI as if we all were researchers at SV ^^

  • @RockEsper
    @RockEsper Рік тому

    I love learning about AI and your videos are a great way to do that. Subbed.

  • @cjp5690
    @cjp5690 Рік тому

    this is probably one of the most interesting videos ive ever seen....I cant believe it....im living in the future now

  • @lukidjano
    @lukidjano Рік тому +3

    RIP Gordon Moore, right when things started to really heat up he took his leave

  • @davidpacheco5501
    @davidpacheco5501 Рік тому +1

    How do you access the vision component of GPT-4? Is it through the API? Would love to start using it to analyze technical datasheets.

  • @toddwmac
    @toddwmac Рік тому +1

    Great stuff as always. Thank you. I'd definitely take the bet that the rapid advances w/in AI, paired with material science advances, will no doubt accelerate the move towards robotics and especially soft robotics. It also seems likely that the same human style interaction which catapulted Chat GPT into the public discourse, will pave the way for a cobot revolution of sorts. """ For anyone interested in some of those materials, a Bing chat prompt: What recent material science advances are enabling tissue like actuators, joints and sensors in robots

  • @LiveType
    @LiveType Рік тому +2

    The arms race has started with the release of gpt-4. Going to be an interesting year.

  • @GabrielVeda
    @GabrielVeda Рік тому +1

    Thanks for another great update. Sam asked Lex a very interesting question: “Would you open source GPT-4?”. Would you?

  • @Madlintelf
    @Madlintelf Рік тому +1

    Been following the LLMs closely since they were released, and yes it's a convergence in technology. As you said when all these technologies synergize we're going to see quantum leaps in tech. I'm excited, but also a bit apprehensive. No, they aren't intelligent, but it's us humans and the ability to exploit anything we can that worries me. Time will tell, and you have another subscriber. Thanks!

  • @melvingeraldsy1552
    @melvingeraldsy1552 Рік тому

    Always great work from you! Please make a video about the language model that reflects on itself. Have you seen that paper?

  • @rickandelon9374
    @rickandelon9374 Рік тому

    Great video as always. The time and research you put into each one of these is insane!! ❤

  • @user-dz9vg9jw5z
    @user-dz9vg9jw5z Рік тому

    It reminds me of a perspective that: visual information is much denser than that in text , after GPT was trained with image , I believe it will make progress faster than ever.

  • @BiroZombie
    @BiroZombie Рік тому +1

    I'm really amazed by the Conformer Model. It's much better than Google's auto-generated captions. With the Conformer Model, the translated audio has correct spelling, grammar, punctuation, and paragraphs. When you play the audio back, each word is highlighted just like Google's auto-generated text. It's so good at fixing mistakes that you might as well use it and let it fix the errors on its own.

  • @PeterRichardsandYoureNot
    @PeterRichardsandYoureNot Рік тому

    Great…..add eyes. The Forbin Project is coming to life faster than we believed possible.

  • @phen-themoogle7651
    @phen-themoogle7651 Рік тому +3

    Any guesses for when they will release the vision stuff to other paying members? I'm really looking forward to it

  • @stephenrodwell
    @stephenrodwell Рік тому

    Thanks, great content! 🙏🏼

  • @TeamELEMENT115
    @TeamELEMENT115 Рік тому

    Hi, love your content, could you recommend a number of essays that are great for introducing the topic of AI, how it works, etc. to laymen that are attempted to understand the technicalities of the field?

  • @9600bauds
    @9600bauds Рік тому +1

    finally we can have automatic captions that don't suck

  • @Gurzbujar
    @Gurzbujar Рік тому

    I would appreciate a channel video where you explain a bit about yourself and your interest in AI. Otherwise loving all your content keep it up!

  • @TheNeFX
    @TheNeFX Рік тому

    Love your videos. Great Job.

  • @Tahazif_TheCool22
    @Tahazif_TheCool22 Рік тому +1

    great vid as always

  • @neilclay5835
    @neilclay5835 Рік тому

    Nice work dude

  • @smudgepost
    @smudgepost Рік тому +1

    Very interesting! I'm already working on legal documentation and standards. The challenge is the rules need to be proven as an absolute and multidimensional arrays need to evidence lossless answers. You could argue expert systems are needed to manage the data, but an unsupervised neural network can better map non-linear data and use the output for a more structured stump when working with CPA and gradient boosting. The challenge is to reduce pdf > xml/JSON/BSON to an automated process and conserve enough context without losing the integrity of the subject. There's a lot of work to do with paragraph and kernel based vectorization/lemmatization.

  • @DG123z
    @DG123z Рік тому

    Thank you for your amazing videos ❤

  • @JOZZYmusic
    @JOZZYmusic Рік тому +3

    "it had trouble recognizing ramen"
    this will be the key to survival during the ai uprising. We'll all have to wear ramen ghillie suits.

  • @krishp1104
    @krishp1104 Рік тому +1

    As a computer scientists, channels like yours and the advancements in AI have really made me consider getting a masters in AI

    • @robinspanier7017
      @robinspanier7017 11 місяців тому

      do you think you need a master in order to use and develop it? (i am realy curios, as a software architect. my method is usually: lets throw something at the wall first and see what sticks)

    • @krishp1104
      @krishp1104 11 місяців тому

      @@robinspanier7017 Most AI job postings ask for research or experience in the field or a masters/phd

  • @dqieu
    @dqieu Рік тому +1

    Your videos have re-ignited my passion for tech, thank you so much. I am researching ways we could broaden our understanding of consciousness, thus developing better tests and understanding GPT. Do you have any insights?

  • @ukeedge2761
    @ukeedge2761 Рік тому

    Haha love the futurama reference, bet all those older models are going to need an upgrade for compatibility issues. 😂

  • @BarnStangz
    @BarnStangz Рік тому

    Thank you for the video! It's nice to have a channel that gets right to the good stuff. I feel like AI is moving so lightning fast, it's hard to keep up and this channel seems to do a great job of getting right into the good stuff, so that's really appreciated. I really hope that AI is going to make things easier on humans in general, it would be nice to work less and enjoy life more. I can't wait to see where things are in the next 5 years, it's going to be insane!

  • @amintourism567
    @amintourism567 Рік тому

    Thank you. Nice video

  • @andrewdunbar828
    @andrewdunbar828 11 місяців тому +1

    Whether the word "broken" gets breaked, time has already telled.

  • @MegaMichael122
    @MegaMichael122 Рік тому +1

    I think it would be pretty rad if they were to put a powerful ai into a bunch of robots, walking ones, drones, water bots etc that then can go fly around and do research on for example wildlife in jungles and what not. They can be equipped with all different types of sensors and various visions(infra red and what have you)

  • @txuguh
    @txuguh Рік тому

    Keep up the good work 😊

  • @Arnoldismouldy
    @Arnoldismouldy Рік тому +1

    Have you gotten access to plugins yet? Ill be waiting for that vid

  • @adisharr
    @adisharr Рік тому

    I think I'm with many others in feeling a sense of amazement and dread at ho the upcoming AI models will change things. Eventually I doubt we'll understand how any of it works past a point.

  • @SoftwareSavant
    @SoftwareSavant Рік тому

    I'm really excited about what AI can do in the medical field. I was just telling my co-worker on Friday about The effects of palm AI and what that may Mean for the medical profession coming soon.

  • @ondrejmarek1980
    @ondrejmarek1980 Рік тому +1

    can you do a deeper dive on that Lex-Sam interview? there was a lot more said and discussed that would be imo worth analyzing more

  • @evreninanc
    @evreninanc Рік тому

    Great video as always. How long do you reckon until we’ll be able to create our digital replicas by telling our life stories, memories and thoughts to our own AI’s?

    • @aiexplained-official
      @aiexplained-official  Рік тому

      Fairly soon in basic terms. In more detailed terms with face mapping etc maybe a year or two.