A lot of smart folks working on leading AI models are never able to tell you specifically what kind of new (large volume) jobs will be created by AI is coz they intuitively know those will also be performed by AI eventually.
Ask anyone: @sama@elonmusk@demishassabis Jensen @DarioAmodei
Itโs not surprising the leading AI companies hiring aggressively but to create tools that automate more and more to reduce more and more jobs.
Productive tools increase jobs yes but automation reduces them.
At this point we should have really identified which *new* jobs are getting created by AI.
A lot of smart folks working on leading AI models are never able to tell you specifically what kind of new (large volume) jobs will be created by AI is coz they intuitively know those will also be performed by AI eventually.
Ask anyone: @sama@elonmusk@demishassabis Jensen @DarioAmodei
AI is the *tool*
Let me explain -
Current state of AI requires it to be integrated with existing tools to power them to the next level.
The tools add little value and AI does the power lifting.
Future versions of AI will achieve what these tools do as well and replace them.
@benitoz He absolutely has no clue or imagination about how much inference the world is gonna need in time to come. He is not alone though. Many people canโt grasp exponentials.
@BradleyKellard He missed the Nvidia rally big time and also sold off too early. Itโs 50% up from last time he spoke about Nvidia monopoly not being durable in mid 2024.
@amitisinvesting AI itself is the moat for Nvidia. We are still at level 1 in a multilevel, multi decade AI game. AI is so powerful that it is threatening the durability of google as well.
AI is the *tool*
Let me explain -
Current state of AI requires it to be integrated with existing tools to power them to the next level.
The tools add little value and AI does the power lifting.
Future versions of AI will achieve what these tools do as well and replace them.
AI is the *tool*
Let me explain -
Current state of AI requires it to be integrated with existing tools to power them to the next level.
The tools add little value and AI does the power lifting.
Future versions of AI will achieve what these tools do as well and replace them.
AI is not software. Hence lot of people donโt have the natural intuition for it similar to how lot of non software people donโt have it for software.
For software, people miss things like scalability, replicability, compute speed, marginal cost etc.
Similarly for AI, people (including from software) miss things like semantic search, one shot intelligence, agentic behavior etc.
People are severely underestimating the data flywheel effect and self recursion moats the frontier AI models have by now. Gonna be extremely difficult to break away from that.
Data Flywheel: A self-reinforcing loop where more usage of an AI model automatically generates data that makes the model smarter, which attracts even more usage this creating compounding intelligence and a widening competitive moat.
AI is the *tool*
Let me explain -
Current state of AI requires it to be integrated with existing tools to power them to the next level.
The tools add little value and AI does the power lifting.
Future versions of AI will achieve what these tools do as well and replace them.
@chamath To be fair these are physical jobs and he often mentioned robotics is still further away. He has always been clear about knowledge work jobs getting disrupted.
AI is the *tool*
Let me explain -
Current state of AI requires it to be integrated with existing tools to power them to the next level.
The tools add little value and AI does the power lifting.
Future versions of AI will achieve what these tools do as well and replace them.
People are severely underestimating the data flywheel effect and self recursion moats the frontier AI models have by now. Gonna be extremely difficult to break away from that.
Data Flywheel: A self-reinforcing loop where more usage of an AI model automatically generates data that makes the model smarter, which attracts even more usage this creating compounding intelligence and a widening competitive moat.
"The fear is that weโre in a similar situation to the tech bubble in 2000-2001, or the railroads in the late 19th โ early 20th century. In both cases, the investment (in Internet cables, websites, railroads) was amazing in the long term, but there was initial overinvestment."
@paulg@Noahpinion IMO this is more of a startup vs big successful company hiring analogy. US is no longer a startup and hence would rather recruit only the best after intense scrutiny.