Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenAI Jalapeño: Better than Nvidia Blackwell

newsletter.semianalysis.com

Post reply on HN
Print view View original

231–240 of 389 posts

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#231

Post by Eisenstein »

Ductapemaster wrote: post #226
Eisenstein wrote: post #145

Earlier quoted context omitted.

The system which runs coolant over the chips can be closed but the part which uses an evaporative system to cool that is still open loop and vents water into the air, no?

Nope, it doesn't have to be open loop! Example of such a system being used specifically for datacenters: https://blog.vantage-dc.com/2026/04/22/cooling-without-the-d... Evaporated water is condensed, and in the process transfers its heat into another place that removes it. Another simple example is a pot of boiling water with a lid on it.

The link you cited is not evaporative cooling and a pot of boiling water with a sealed lid on it is a pressure vessel which eventually explodes.

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#232

Post by Ductapemaster »

Eisenstein wrote: post #231
Ductapemaster wrote: post #226

Earlier quoted context omitted.

Nope, it doesn't have to be open loop! Example of such a system being used specifically for datacenters: https://blog.vantage-dc.com/2026/04/22/cooling-without-the-d... Evaporated water is condensed, and in the process transfers its heat into another place that removes it. Another simple example is a pot of boiling water with a lid on it.

The link you cited is not evaporative cooling and a pot of boiling water with a sealed lid on it is a pressure vessel which eventually explodes.

If you were to remove the heat at a sufficient rate by, say, turning the lid into a heat exchanger, you would have a stable system.

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#233

Post by undersuit »

falcor84 wrote: post #202
cmrdporcupine wrote: post #164

Earlier quoted context omitted.

Right, I can do the talked about ~3 tok/sec output and drive a car, hold my bladder, and eat chips at the same time. Take that, Jalapeno!

For what it's worth, LLMs don't really suffer from incontinence, so at least that part is pretty much a solved problem.

So why are their water cooling systems filled with leak detectors? /s

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#234

Post by newyankee »

mkw5053 wrote: post #150

Warning, this is a long comment! (I’m trying to stick to sourced facts here and not overstate what they mean) I went down a rabbit hole after watching Dylan Patel on Dwarkesh today: https://www.youtube.com/watch?v=aV26V1UvkJw I was initially just surprised by how bullish Dylan is on OpenAI/Anthropic and how bearish he is on China, despite Chinese labs getting closer to US SOTA while offering inference at dramatically…

If you research the origins of Dwarkesh, even more conspiracy level points emerge. I do not think even in the handful videos post Leopold fund collapse he has addressed it in any ways. That is the point of so called observers, they pretend to be impartial but everyone is a hustler in some way.

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#235

Post by Eisenstein »

Ductapemaster wrote: post #232
Eisenstein wrote: post #231

Earlier quoted context omitted.

The link you cited is not evaporative cooling and a pot of boiling water with a sealed lid on it is a pressure vessel which eventually explodes.

If you were to remove the heat at a sufficient rate by, say, turning the lid into a heat exchanger, you would have a stable system.

How is that different from not using evaporative cooling and just putting the heat exchanger on the burner?

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#236

Post by smallmancontrov »

kushie wrote: post #98
bmulholland wrote: post #97

Earlier quoted context omitted.

Probably! But not viable yet; the chips would be about a year behind SOTA. Note the ~16 months that the article quotes as being insanely fast to get this chip to tape-out (read: start producing). We'll have to bootstrap our way there: AI is actively being used to get us closer to viable lead times for this. Unfortunately, there's some real physical constraints: IIRC, manufacturing a wafer takes on the order of a mont…

tapeout could shrink but days per mask layer (DPML) does not have much margin..

I'm not in industry, is DPML (which I assume is the time required to make a mask?) set by electron beam scan time or something?

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#237

Post by guhcampos »

mf_tomb wrote: post #190
mchusma wrote: post #85

I think they talked about this being general purpose chip but I would think that Anthropic/OpenAI are at the scale now they could bake LLM weights into chips themselves. For example, GPT Sol baked into a custom chip run for $100M that runs 10x as fast and 10x as cheap should pay for itself as long as the chip is useful for long enough. While 2 years ago nothing was useful more than 1 year long, there are many older m…

"Baking in" a model into a chip is a bad idea because chips take 2 years to tape out and then you're stuck doing inference on llama 3 in 2026 when fable/sol are available. Every accelerator is a tradeoff between flexibility and performance and GPUs are already pareto-optimal

There are other options. I worked for a startup called NVXL and we were programming DNNs into FPGAs using OpenCL, on custom boards we built to plug into NVME. It worked great, but it wasn't fast enough at the time to compete with Nvidia, or even Intel AVX512. Ultimately the company failed, but maybe some hybrid like that could work for LLMs? I haven't been up to date on how DNNs and LLMs look like under the hood these days, but there's got to be someone doing something similar.

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#238

Post by MBCook »

kurthr wrote: post #161
bmulholland wrote: post #97

Earlier quoted context omitted.

Probably! But not viable yet; the chips would be about a year behind SOTA. Note the ~16 months that the article quotes as being insanely fast to get this chip to tape-out (read: start producing). We'll have to bootstrap our way there: AI is actively being used to get us closer to viable lead times for this. Unfortunately, there's some real physical constraints: IIRC, manufacturing a wafer takes on the order of a mont…

The metal masked ROM is basically only 2 metal/contact layers. It's not a full new design and tapeout. You could roll a new set of parameters every ~2-3months. It's not an architectural change. See statements below. https://www.eetimes.com/taalas-specializes-to-extremes-for-e... https://www.turingpost.com/p/taalas https://cambrian-ai.com/taalas-launches-hardcore-chip-with-i... Part of the key is that by moving even f…

But that means your different chips all have different sets of weights and are different generations.

If none of that is baked into the chip as now then all the chips are running the latest weights every time.

Even if you could ignore the stuff built into the chip when the time came, at that point you just wasted money on silicon that’s useless in 2-3 months.

Re: OpenAI Jalapeño: Better than Nvidia Blackwell

#239

Post by onion2k »

rustystump wrote: post #160
LogicFailsMe wrote: post #152

Earlier quoted context omitted.

Along with Leopold Aschenbrenner so maybe not so much.

The guy that was part of FTX, fired from openai for alleged theft, got billions in a hedge fund somehow then lost billions. Why are all these people so scummy? It is like voting Trump three times in a row.

They're very intelligent people who do very clever things at a young age, which draws the attention of very rich people who can exploit them to get richer, and no one tells the young person they're being exploited. They're heaped with praise and 'wealth' (millions, but crumbs compared to what they're making for other people), and told they're geniuses who can do no wrong, mostly by the media that happens to be owned by the rich.

Then the rich people pull the rug leaving them holding the bag, and they move on to the next young clever group.

And the cycle continues.

Post reply on HN