Post

Slow down frontier model development. Bring open weights to current frontier. Build out local inference providers, stabilize hardware prices, focus on cost and token throughput for a bit. Get our software updated to handle what's being thrown at us. I would not mind spending a beat on distribution.

Do you receive financial or growth forecasts and wonder why they think exponential QoQ growth is possible? This will be a rug-pull. Hype the stock, promise deals to one another for future hardware purchases to muddy the actual profits and hope the bubble doesn't pop until after milking govt funds.

I think the frontier labs didn't expect their work to be so easily commoditized

But the people betting on valuation do.

bag holder politics

Tulip mania seems to be a cyclical thing. .com bust had very similar patterns of over-hype of the value 10-15 years before it was viable. I think the challenge is that we allow skeptics and writers who do hard research to be written off as FUD while NYT does puff pieces on the Theranos of the day.

yeeeeah I think crypto was more tulip mania. this is more like railroad bubble. completely real, difficult to finance with the right sequence of returns

Possibly, but the circular nature of investment and threat of losing their infinite water access, priority power access and lack of property taxes in places like Texas makes the RoI probability even lower than it is currently.