1. Forum
  2. Newsgroups
  3. Alt.2600

  • Has Open Source AI Just Become Too Dangerous?

    From rek2 hispagatos@rek2@hispagatos.org.invalid to alt.2600.madrid,hispagatos.talk,alt.2600,alt.2600.madrid on Wednesday, July 31, 2024 20:36:56
    From Newsgroup: alt.2600

    Very good thoughts.

    Has Open Source AI Just Become Too Dangerous? https://invidious.fdn.fr/watch?v=RhsKgvGue0w

    Did Mark Zuckerberg make a mistake to open source an LLM with 405 billion parame
    ters? Support me: www.patreon.com/thehatedone

    There is a new open source AI out there. It was published by Meta and it’s one
    of the most powerful models ever developed. And now a lot of people want to sto
    p it. They say Meta is too dangerous and irresponsible. And what if they are rig
    ht this time? Why would Mark Zuckerberg do this? Who should you believe and what
    if open source AI is a mistake?

    Let me shine a light on all of these things that I’ve been following extremely
    closely as I believe what will transpire over the next months and years will pr
    ofoundly affect all of our digital lives.

    On July 23, Meta published a new stack of AI models as Llama 3.1 packaged into t
    hree different sizes. There are two models small enough to run on consumer hardw
    are with just 8 and 70 billion parameters. But we’ve had those for a while. Wh
    at really changes everything is Meta’s largest model to date – with whopping
    405B paramaters, it’s the biggest and most capable open source LLM ever creat
    ed.

    The 405b model is now capable of rivaling and even outperforming ChatGPT. No ope
    n source AI has done this before. But unlike GPT-4, anyone in the world can copy
    , modify and redistribute Meta’s flagship AI without almost any restrictions.

    Sources [references available in the transcript]
    [0] ai.meta.com/blog/meta-llama-3-1/
    [1] huggingface.co/meta-llama/Meta-Llama-3.1-405B
    [2] www.ft.com/content/a09e4aaf-be52-4a45-86a7-c6d1636…
    [3] time.com/7002563/mark-zuckerberg-ai-llama-meta-ope…
    [4] www.engadget.com/llama-31-is-metas-latest-salvo-in…
    [5] www.politico.com/news/2023/12/03/congress-ai-fello…
    [6] www.politico.eu/article/rishi-sunak-artificial-int…
    [7] www.afr.com/technology/google-brain-founder-says-b…
    [8] www.ft.com/content/2dc07f9e-d2a9-4d98-b746-b051f93…
    [9] carnegieendowment.org/2023/09/14/how-hype-over-ai-…
    [10] www.msn.com/en-us/news/technology/meta-llama-3-1-i…
    [11] cdn.openai.com/papers/gpt-4.pdf
    [12] arstechnica.com/information-technology/2024/07/the…
    [13] www.latimes.com/business/story/2024-07-23/zuckerbe…
    [14] www.wired.com/story/meta-ai-llama-3/
    [15] fortune.com/2024/07/23/meta-new-llama-model-3-1/
    [16] gs.statcounter.com/os-market-share/
    [17] source.android.com/
    [18] www.axios.com/2022/09/14/google-loses-appeal-eu-an…
    [19] www.npr.org/2023/09/12/1198558372/doj-google-monop…
    [20] about.fb.com/news/2024/07/open-source-ai-is-the-pa…
    [21] blog.mozilla.org/en/mozilla/introducing-mozilla-ai…
    [22] www.weforum.org/agenda/2023/12/ai-regulation-open-…
    [23] www.technologyreview.com/2024/03/25/1090111/tech-i…
    [24] spectrum.ieee.org/open-source-ai-good
    [25] not-just-memorization.github.io/extracting-trainin…
    [26] arxiv.org/abs/2310.15469
    [27] arxiv.org/abs/2310.03693
    --
    - {gemini,https}://{,rek2.}hispagatos.org - mastodon: @rek2@hispagatos.space
    - [https|gemini]://2600.Madrid - https://hispagatos.space/@rek2
    - https://keyoxide.org/A31C7CE19D9C58084EA42BA26C0B0D11E9303EC5
    --- Synchronet 3.20a-Win32 NewsLink 1.114