AI is still a fad.
Only if you’re moving the goal posts every other week.
Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
91–100 of 138 posts
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#92A year ago Sam Altman was going around trying to convince people we all needed to drop 7 trillion dollars to build hundreds of fabs and nuclear power plants to fuel his AI ambitions. Only a week ago he was triumphantly announcing 500 billion dollar deals with our new President. The (regrettably temporary) ousting of Sam Altman looks like the right call, in hindsight. Of course some amount of showmanship is expected,…
The bitter lesson was never about hardware scaling being the only ingredient. It was about favouring generalist approaches that can scale with your hardware budget over carefully handcrafted bespoke solutions that conserve the "cheap" resource.
I'm not seeing anyone at OpenAI abandon the static weights model and yet they have audacity to claim that they just need to scale more?
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#93DeepSeek has demonstrated that there is no technical moat. Model training costs are plummeting, and the margins for APIs will just get slimmer. Plus model capabilities are plateauing. Once model improvement slows down enough, seems to me like the battle is to be fought in the application layer. Whoever can make the killer app will capture the market.
Something worth noting is that ChatGPT currently is the killer app -- DeepSeek's current chart-topping app notwithstanding (not clear if viral blip or long-term trend).
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#94Earlier quoted context omitted.
Honestly, I’m not sure I’m completely sold on the value of LLMs long term but this is the most realistic and reasonable take I’ve read on this post so far. If anything, it’s an downward adjustment in the cost implications but could actually unlock exponential improvements on a shorter time horizon than expected because of that. Investors getting scared probably is a good opportunity to buy in.
Bullish on the use. Bearish on the profit margins for the big players. If (big if!) I understand correctly, the ceiling for edge/local/offline AI has just blown off.
Is there an acronym for edge/local/offline? ELO could be confused with something AI already dominates at. As someone working in the edge/local/offline space it’s interesting to hear these together though. Offline is local but local often isn’t offline :)
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#95Earlier quoted context omitted.
It's always been possible to "do (worse) AI on less compute". We've had years of open models! I also don't understand how anyone can see this as anything but good news for OpenAI. The ultimate value proposition of AI has always depended on whether it stretches to AGI and beyond, and R1 demonstrates that there's several orders of magnitude of hardware overhang. This makes it easier for OpenAI to succeed, not harder, b…
The interesting part is that distillations based on reinforcement learning based models are performing so well. That brings the cost down dramatically to do certain tasks.
I thought the distillations were SFT only?
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#96Earlier quoted context omitted.
Honestly, I’m not sure I’m completely sold on the value of LLMs long term but this is the most realistic and reasonable take I’ve read on this post so far. If anything, it’s an downward adjustment in the cost implications but could actually unlock exponential improvements on a shorter time horizon than expected because of that. Investors getting scared probably is a good opportunity to buy in.
Bullish on the use. Bearish on the profit margins for the big players. If (big if!) I understand correctly, the ceiling for edge/local/offline AI has just blown off.
Bullish on the prospects for small players, then.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#97What a great take, I have thought this for a while.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#98Earlier quoted context omitted.
I thought Nvidia is selling shovels
Mickey Mouse animated the shovels with the big spell book, and now they're marching around and need shovels of their own...
This has me imagining it in terms of the game Cuphead. That’s also good for visualizing this I think.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#99A year ago Sam Altman was going around trying to convince people we all needed to drop 7 trillion dollars to build hundreds of fabs and nuclear power plants to fuel his AI ambitions. Only a week ago he was triumphantly announcing 500 billion dollar deals with our new President. The (regrettably temporary) ousting of Sam Altman looks like the right call, in hindsight. Of course some amount of showmanship is expected,…
Latest word from The Leader: "Even as some U.S. tech stocks plunged on Monday after it appeared that DeepSeek could produce similar results as rival models with a system that was cheaper to build, Trump projected confidence, calling it “very much a positive development.” He reasoned that American companies would be able to adapt and evolve based on DeepSeek’s demonstration that effective systems can be developed more…
It’s not wrong
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#100> But while Facebook’s costs decreased as it scaled, OpenAI’s costs are growing in lockstep with its revenue, and sometimes faster And here comes DeepSeek and takes the steam out of this and the cost arguments that follow it.
It's lose-lose for their valuation regardless. This scenario might if anything be worse for them. Now they have massive sunk capital investments that the second mover might be able to avoid. If the open source models get small enough and high enough quality, the rationale for runing them in the cloud in the first place start to evaporate. How does OpenAI get paid for a use case that can easily be run locally on an iP…
Because they deliver models you cannot run locally on an iPhone, and those are almost always the models you want to run.