• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Pranav Shyam Retweets Andrew Ho Reply

    Google DeepMind researcher shares @andrewho03 reply on model observations.

    OK
    JB
    PS
    7 Sources, 27d ago, first seen 27d ago

    TLDR

    Pranav Shyam, identified as a research scientist at Google DeepMind with prior work on generative models at OpenAI, retweeted a post by @andrewho03. That post replies to @AgustinLebron3 and opens by stating the observations are not new to the author. It begins a list with the phrase "Despite the seemingly magical nature of" before the visible text cuts off. The packet records only this retweet action and the quoted opening lines among visible posts.

    Combined views

    351.4K

    7 Sources, first seen 27d ago

    Combined views

    351.4K

    7 Sources, first seen 27d ago

    3.2K likes
    3.2K likes
    102 comments
    1.8K saves
    484 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    102 comments
    1.8K saves
    484 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    7 Sources

    @andrewho03Not really observations that others before me haven’t made, but: - Despite the seemingly magical nature of LLMs, reflection over a >3 month timescale suggests my total productivity hasn’t increased by over 100%, or perhaps even by over 50%, and a lot of time is actually wasted because LLMs enable me to spend time on gratifying but low-productivity tasks that in the future turn out to not be useful - Also, capabilities are incredibly spiky and highly correlated with the degree of investment poured into them, which my earlier tweet about math benchmarks implicitly points out - From the above, it seems that the nature of LLM intelligence is wildly dissimilar to that of human intelligence and we won’t trivially get to something superior to human intelligence in all important respects just by scaling up existing approaches with various tweaks; even if AGI Is eventually achievable, this implies a significantly longer timeline - Benchmark progress is almost definitionally guaranteed to happen because the process of constructing a benchmark is a direct precursor to the process of constructing a training dataset used for hill climbing that benchmark, but the scope of what can be captured in a benchmark is (at least for now) grossly lacking in terms of its relevance to real-world work, with maybe several limited exceptions - Progress seems highly gated by data but the nature of model training means that each “next dataset” is significantly harder to assemble than what preceded it; some wins are possible through synthetic methods but those feel more like “patching up gaps” than “pushing the frontier forward” At a higher level, I guess I’d say there’s a sort of refusal to think carefully about what models are or are not useful for in a rigorous way which I find personally quite annoying, and instead a reliance on some nebulous notion of being “AGI pilled” as a replacement for serious thought. I think people are very quick to anthropomorphize LLM intelligence because humans communicate through words and we infer the intelligence of human counterparties through comprehension of their language, but this leads them to wrong conclusions; for example if we observe that a new model proved some incredible mathematical theorem, some will say, “well, don’t we have AGI now, huh?” But to me, it’s actually more like, “well, given how hard it would have been for a human to do these mathematics, and given the limited economic effect of LLMs upon the world so far, isn’t it actually a negative datapoint vis-a-vis the generality of LLM intelligence?”
    @recurseparadoxRT @andrewho03: @AgustinLebron3 Not really observations that others before me haven’t made, but: - Despite the seemingly magical nature of…
    @lateinteractionRT @andrewho03: @AgustinLebron3 Not really observations that others before me haven’t made, but: - Despite the seemingly magical nature of…
    @hsu_steveI find the models very impressive bc they improve my productivity in math-phys research. Most SWEs are similarly impressed. But these are areas the labs specifically targeted - spending billions on data generation (both synthetic and from human experts), complex training environments, etc. I'm perhaps more optimistic than Andrew (below) about how easily these capabilities generalize to areas the labs have not specifically targeted. However, productivity impact in the general economy may take years to materialize. Progress will be slowed by 1. human gatekeeping and 2. in-the-weeds challenges of figuring out how best to apply alien machine intelligence. This applies to most novel technologies or tools. A related point I sometimes make: look at all the Fields Medal results of the last decades. How much GDP impact has it made? Almost none - certainly not detectable in economic statistics. Yet, those people are geniuses - at the pinnacle of human intelligence. There is clearly a distinction to be made btw "very nontrivial stuff; beyond the capability of almost all humans" and "this improves total factor productivity"! I think the models are capable of the latter (SWE, chip design, medical diagnosis, science research), but it's a nontrivial challenge to make it happen.
    @PlinzRT @hsu_steve: I find the models very impressive bc they improve my productivity in math-phys research. Most SWEs are similarly impressed.…

    7 Sources

    @andrewho03Not really observations that others before me haven’t made, but: - Despite the seemingly magical nature of LLMs, reflection over a >3 month timescale suggests my total productivity hasn’t increased by over 100%, or perhaps even by over 50%, and a lot of time is actually wasted because LLMs enable me to spend time on gratifying but low-productivity tasks that in the future turn out to not be useful - Also, capabilities are incredibly spiky and highly correlated with the degree of investment poured into them, which my earlier tweet about math benchmarks implicitly points out - From the above, it seems that the nature of LLM intelligence is wildly dissimilar to that of human intelligence and we won’t trivially get to something superior to human intelligence in all important respects just by scaling up existing approaches with various tweaks; even if AGI Is eventually achievable, this implies a significantly longer timeline - Benchmark progress is almost definitionally guaranteed to happen because the process of constructing a benchmark is a direct precursor to the process of constructing a training dataset used for hill climbing that benchmark, but the scope of what can be captured in a benchmark is (at least for now) grossly lacking in terms of its relevance to real-world work, with maybe several limited exceptions - Progress seems highly gated by data but the nature of model training means that each “next dataset” is significantly harder to assemble than what preceded it; some wins are possible through synthetic methods but those feel more like “patching up gaps” than “pushing the frontier forward” At a higher level, I guess I’d say there’s a sort of refusal to think carefully about what models are or are not useful for in a rigorous way which I find personally quite annoying, and instead a reliance on some nebulous notion of being “AGI pilled” as a replacement for serious thought. I think people are very quick to anthropomorphize LLM intelligence because humans communicate through words and we infer the intelligence of human counterparties through comprehension of their language, but this leads them to wrong conclusions; for example if we observe that a new model proved some incredible mathematical theorem, some will say, “well, don’t we have AGI now, huh?” But to me, it’s actually more like, “well, given how hard it would have been for a human to do these mathematics, and given the limited economic effect of LLMs upon the world so far, isn’t it actually a negative datapoint vis-a-vis the generality of LLM intelligence?”
    @recurseparadoxRT @andrewho03: @AgustinLebron3 Not really observations that others before me haven’t made, but: - Despite the seemingly magical nature of…
    @lateinteractionRT @andrewho03: @AgustinLebron3 Not really observations that others before me haven’t made, but: - Despite the seemingly magical nature of…
    @hsu_steveI find the models very impressive bc they improve my productivity in math-phys research. Most SWEs are similarly impressed. But these are areas the labs specifically targeted - spending billions on data generation (both synthetic and from human experts), complex training environments, etc. I'm perhaps more optimistic than Andrew (below) about how easily these capabilities generalize to areas the labs have not specifically targeted. However, productivity impact in the general economy may take years to materialize. Progress will be slowed by 1. human gatekeeping and 2. in-the-weeds challenges of figuring out how best to apply alien machine intelligence. This applies to most novel technologies or tools. A related point I sometimes make: look at all the Fields Medal results of the last decades. How much GDP impact has it made? Almost none - certainly not detectable in economic statistics. Yet, those people are geniuses - at the pinnacle of human intelligence. There is clearly a distinction to be made btw "very nontrivial stuff; beyond the capability of almost all humans" and "this improves total factor productivity"! I think the models are capable of the latter (SWE, chip design, medical diagnosis, science research), but it's a nontrivial challenge to make it happen.
    @PlinzRT @hsu_steve: I find the models very impressive bc they improve my productivity in math-phys research. Most SWEs are similarly impressed.…