Trillium TPU, built to power the future of AI

Поделиться
HTML-код
  • Опубликовано: 2 янв 2025

Комментарии •

  • @returnedinformation1040
    @returnedinformation1040 2 месяца назад +173

    What about the successor of Google Coral? I'm asking for us mortals

    • @Phil-D83
      @Phil-D83 2 месяца назад +7

      Will be waiting for awhile

    • @BlackDub21
      @BlackDub21 2 месяца назад +2

      @@Phil-D83after google is broken up probably

  • @sharex21
    @sharex21 2 месяца назад +55

    The narration feels like it's been made by a vlogger who just started yesterday in his basement.

  • @joe_hoeller_chicago
    @joe_hoeller_chicago 2 месяца назад +31

    TPUs although very useful and a big gain in speed for training large AI models, are only on GCP. So if you’re not on that cloud, you can’t use them.

    • @NadimShaikh-qv7zj
      @NadimShaikh-qv7zj 2 месяца назад +8

      Yeah, that's the entire business proposition of this product. They want to sell you their cloud services, not their AI hardware per se.

  • @Atom224
    @Atom224 2 месяца назад +16

    Will there be special programs/offers for university research institutes?

    • @lgmuk
      @lgmuk 2 месяца назад +2

      There's one already! Check the Gemma documentation page

  • @nike5428
    @nike5428 Месяц назад +2

    The Background music sometimes is greater than or equal to the voice of the narrator, making it difficult to understand.

  • @AlienAnthony
    @AlienAnthony Месяц назад +3

    When I can buy the hardware, I will be interested.

  • @arirajuns
    @arirajuns 2 месяца назад +5

    Thanks for the notebookllm. Its very useful.

  • @WhyInnovate
    @WhyInnovate 2 месяца назад +3

    So little details, what it’s tflops at 32,16,8 bits, what is optimized for only tensor math or LLM inference? So many questions

  • @HeavenlyPalace-Y2K
    @HeavenlyPalace-Y2K Месяц назад

    How's the performance compared to Nvidia? I guess there are no easy benchmark comparisons out there like PC hardware performance.

  • @user-w9qcxxl5va
    @user-w9qcxxl5va Месяц назад +1

    와 미친 이 광고영상 AI로 생성한거네 ㅋㅋㅋㅋㅋㅋㅋㅋ

  • @synthwave7
    @synthwave7 2 месяца назад +1

    Google and Microosoft leading in AI architecture.

    • @DeathHeadSoup
      @DeathHeadSoup Месяц назад +1

      Nah, neither of those companies are building neuromorphic architectures, they are actually way behind in that regard. They will probably just buy a startup like Rain Neuromorphic or BrainChip Holdings Ltd. IBM and Intel are doing fairly well in the neuromorphic space.

    • @deeplearningpartnership
      @deeplearningpartnership 22 дня назад

      Microsoft?

  • @obviouswarrior5460
    @obviouswarrior5460 Месяц назад

    My NVR need 2-3 Pcie card of that !
    How much ? Where to buy ?

  • @Unineil
    @Unineil 2 месяца назад +1

    I want a tpu on my pi that interacts more with local 1-3b llms.

  • @KevinKreger
    @KevinKreger Месяц назад

    It's not like they won't charge 4.7 times as much to use them.

  • @cystarkman
    @cystarkman 2 месяца назад

    That facility reminds of other earlier and also larger platforms for intelligence.

  • @Aarti-c4r
    @Aarti-c4r 2 месяца назад +2

    Great announcement 💥💯

  • @julioconradomarinardila3269
    @julioconradomarinardila3269 2 месяца назад

    Excelente herramienta para LA Google cloud

  • @ejaygerald7877
    @ejaygerald7877 2 месяца назад

    TPUs. Mga gawa ko yan, gawa namin yan.
    Google is mine.

  • @rowandevving
    @rowandevving Месяц назад

    Yay! I can now use 100 more megawatts of power to generate marginally better AI slop at scale!

  •  2 месяца назад +4

    Will it be used in Project Nimbus in the West Bank?

    • @BlackDub21
      @BlackDub21 2 месяца назад

      Will 😂 try something past tense.

  • @CodecookerMintz
    @CodecookerMintz Месяц назад

    So, Can we use it for free in google colab

  • @ADVwilly
    @ADVwilly 2 месяца назад

    But will it run Crysis?

  • @andreamanninfiaschi1679
    @andreamanninfiaschi1679 2 месяца назад +2

    All of this power, all of this technology...for watching dogs go surfing.

  • @MichaelSkinner-e9j
    @MichaelSkinner-e9j Месяц назад

    I wonder what AMD and Nvidia's answer to this is

  • @themozesmugagga
    @themozesmugagga Месяц назад

    google just keeps swelling
    :Damn

  • @Michaelno
    @Michaelno Месяц назад

    Google Stadia?! Where you at? 😂

  • @SimonOrlovsky
    @SimonOrlovsky 2 месяца назад +2

    This is awesome. Great presenter.

  • @dineshlamarumba4557
    @dineshlamarumba4557 2 месяца назад

    The future of ai harsware technology should be such that we xan also develop micro intelligent robots, how will gpu, tpu in any way fit such micro robots?

  • @rogerc7960
    @rogerc7960 2 месяца назад

    VDU?

  • @seansingh4421
    @seansingh4421 Месяц назад +2

    Saying a lot of “Keywords” only doesn’t constitute a complete sentence in any language.

  • @Adolphout
    @Adolphout 2 месяца назад

    When is this going to be available on the standard android mobile device?

  • @animationgaming8539
    @animationgaming8539 2 месяца назад

    Nice

  • @mcd123
    @mcd123 2 месяца назад

    Looking good! Accelerate is the way….to 1 Trillium FLOPS!

  • @EarthCreature.
    @EarthCreature. 2 месяца назад +2

    Not 1 solar panel on Google's building. Shame on them

  • @_XY_
    @_XY_ Месяц назад +1

    Matrix

  • @syntaxed2
    @syntaxed2 2 месяца назад +7

    If only the realtime translation was good - It and subtitles are horrible, no one uses it but hey thanks for forcing me to verify login/location every 5minutes

  • @GobalKrishnanV
    @GobalKrishnanV 2 месяца назад

    I have doubt, human brain not have skill all topic , even for single topic no one be excellent. Then how we say want we doing is correct?😅😊❤❤

  • @sahilx4954
    @sahilx4954 2 месяца назад

    No more cpus or gpus for humans anymore. 😢
    Everything is for AI, from now on. 🥺

  • @TaylorSwifty69
    @TaylorSwifty69 Месяц назад

    Hard to understand with this narrator. Just why?

  • @Rr-cr4qu
    @Rr-cr4qu 2 месяца назад

    why choose slowpoke as a dictor

  • @grzegorzmajewski591
    @grzegorzmajewski591 Месяц назад

    Blokujecie GPT AI w Polsce usługi działają tylko w synchronizacji Dysk Google z licencją Major Free Cloud P1 lub P2 z usługą CSS The encryption code TSL Standard Gmail w wersji dla dzieci i młodzieży Account Password Apache 2.0 2013 rok Insert the password Port 0 Hyperlink Hyper-V vs VPN Free Wi-Fi Android 📱 🇵🇱 👎

  • @Tensor-An
    @Tensor-An 2 месяца назад +1

    Game still lags

  • @matt37221
    @matt37221 2 месяца назад

    i think AI is out and too hyped, i am betting bigger on the metaverse and organoid intelligence

  • @HaiLeQuang
    @HaiLeQuang 2 месяца назад

    There is no point to build AI chip if your AI product sucks. Move fast or you would be the Nokia of 2020s years.