kioku

I am slower than a computer at computation, so why would I compete against it?

While I can, at least partially, empathize with the fact that the changing tide is closing in very fast, maybe we as engineers should lean more into what makes us such, and focus on applying principles in order to build machines.

I think it does what it’s supposed to do.

I’m just wondering how this translates to computer manufacturers like Apple. Could we have these kinds of chips built directly into computers within three years? With insanely fast, local on-demand performance comparable to today’s models?

Quite unforgiving. It quickly confirmed the sloppiness of a few titles that are currently sitting on the front page.

> Our key insight is to offload critical softmax primitives to idle tensor units, maximizing hardware utilization and throughput.

> … speedups of 1.05–1.17×across diverse attention configurations on Ampere and Hopper GPUs …

I found the captions on Figure 1 quite interesting.

> Average performance (%) across four agentic benchmarks improves consistently with increasing model Intelligence Index.

> Centralized and hybrid coordination generally yield superior scaling efficiency, suggesting that collaborative agentic structures amplify capability gains more effectively than individual scaling alone.

Then again, the deltas between SAS and best performing MAS approach are ~8%, so I can't help wonder if it's worth the extra cost, at least for the generation of models that was studied.

> This isn't just compliance theater; it's a straight‑up national economic security play.

The woes of LLM contrasts…

In all seriousness, the points made ring true not only for European companies and should make everyone consider the implications of the current situation, as dreary as they are.

Sounds like your experience matches what's being described in the paper. Even if we get correct code, it might not carry design intent.