The paperclip maximizer story, where an AGI directs all economic resources towards manufacturing paperclips, was wrong. We have the the graphics card maximizers, where humans voluntarily redirect vast economic resources towards generating a maximal stream of tokens that hardly anyone wants. Or perhaps call them token maximizers.
There used to be the notion that "talk is cheap". Now we are spending $trillions on generating idle talk.
People continues to mis-weight the benefits of AGI / LLM or in the case "graphics card maximizers". It is precisely because of this, we can continue to extend semi-conductor improvements for at least another 5 years, brings in 1200W CPU in server and liquid cooling along with new DC and Rack design 5 years earlier. Partly continues to fund Gaming GPU where the industry itself were not able to sustain it. New improvement to Networking Gear due to heavy use of LLM. HBM profits that drives current EUV DRAM innovations.
As far as I can tell, I see very little down size to all these investments where Big Tech would have spent those money on stock buy backs or dividends.
And this is assuming AGI really is a bubble and brings little to no productivity, and exclude all the improvements that is adjacent to it.
I'm still not sure why anyone would buy a dell server. Supposedly xAI buys from them, which probably accounts for this, but it is generally a much more sensible choice to buy from companies that are lower-cost and less focused on selling you "enterprise support" (supermicro et al).
How long do you think we have before the AI Bust? This isn’t a rhetorical question, I’m looking for estimates before I have to exit all financial markets and go full defensive in my portfolio. Can it make to 2026? Or will the bust be several years from now? I know no one really knows, but estimates can be very helpful. Sometimes you can feel the wind change.
My company sells them all. Dell, Lenovo, Supermicro, HP (if you insist). I can have anything I want and pretty much for free. Internally we are 90% Dell, our remote employees all have laptops with ProSupport and on-site service because the price is lower than a single day of downtime. They have the infrastructure and logistics to fix anything nearly anywhere. No qualms about recommending them for anything.
11 comments
[ 3.8 ms ] story [ 48.7 ms ] threadThere used to be the notion that "talk is cheap". Now we are spending $trillions on generating idle talk.
People continues to mis-weight the benefits of AGI / LLM or in the case "graphics card maximizers". It is precisely because of this, we can continue to extend semi-conductor improvements for at least another 5 years, brings in 1200W CPU in server and liquid cooling along with new DC and Rack design 5 years earlier. Partly continues to fund Gaming GPU where the industry itself were not able to sustain it. New improvement to Networking Gear due to heavy use of LLM. HBM profits that drives current EUV DRAM innovations.
As far as I can tell, I see very little down size to all these investments where Big Tech would have spent those money on stock buy backs or dividends.
And this is assuming AGI really is a bubble and brings little to no productivity, and exclude all the improvements that is adjacent to it.
It would behoove a lot of these companies. There again say $13 billion series f to make an AWS competition stack.
I mean you have all the infrastructure to magically write the code for you right?