News for hackers. Stuff that matters.
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by slashdave on Sunday May 17, 2026 @11:58AM (#48170074)
We are discussing how rapid development has been, and now you want to freeze your model in silicon?
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by nixon_why69 on Sunday May 17, 2026 @12:07PM (#48170153)
Why not have a bunch of SRAM and various operations like "Q4 matmul" in silicon? Model weights and even architectures could still evolve on a platform like that.
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by ac29 on Sunday May 17, 2026 @12:56PM (#48170664)
Doesnt "a bunch of SRAM" top out at maybe a few gigs per chip (with zero area used for logic)? You'd need an order of magnitude more to fit even a fairly weak general purpose LLM model.
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by throwa356262 on Sunday May 17, 2026 @12:57PM (#48170673)
I belive that is what NPUs are.

The issue is the very huge amount of DRAM and high bandwidth these model require.

Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by rogerrogerr on Sunday May 17, 2026 @11:59AM (#48170086)
Genuine question from a place of ignorance: what in the silicon pipeline makes it take 2-4years to produce chips with a new model on them? Curious what the process bottleneck is.
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by pjc50 on Sunday May 17, 2026 @12:14PM (#48170239)
I think you could get it down to three months between weight changes, if you can encode it in metal layers only. The remaining limits are the fab lead time, and the cost of a metal respin (hundreds of thousands to millions of dollars depending on process).
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by jazzyjackson on Sunday May 17, 2026 @12:08PM (#48170164)
Without being an insider, I imagine that most global fab capacity is contracted out several years in advance.

You might be interested in the tiny tape out project, which guides you through the process of getting your own design etched on silicon. If you only need larger features and not the next gen single digit nanometer stuff, you may not be so supply constrained.

https://tinytapeout.com/

Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by happosai on Sunday May 17, 2026 @02:54PM (#48172059)
I think that comment meant it's 2-4 years until local models are good enough that it's worthwhile to burn an ASIC of them. Not that it takes 2-4 years to make an ASIC chip.
Re:Every AI Subscription Is a Ticking Time Bomb for Enterprise by dangus on Sunday May 17, 2026 @12:12PM (#48170221)
If the silicon costs $200-300 and the company throws it away every two years that’s cheaper than a subscription.

Also, how many companies will just buy an M6/M7 MacBook Pro with 32GB+ of RAM in a couple of years and get “free” AI along with the workstation they were going to buy anyway?