Hacker News: airspresso

New comment by airspresso in "DeepInfer: High-performance inference framework for DeepSeek models"

airspresso — Tue, 20 Oct 2026 20:06:48 +0000

> "A computer can never be held accountable, therefore a computer must never make a management decision."

Legendary quote from an internal 1979 IBM training document [1]

https://simonwillison.net/2025/Feb/3/a-computer-can-never-be...

New comment by airspresso in "Intel D270 GPU AI Accelerator–Up to 480 GB LPDDR5X"

airspresso — Tue, 13 Oct 2026 04:46:28 +0000

Noticed this as well! Was looking for details on the software support and this FAQ entry didn't exactly instill confidence...

New comment by airspresso in "Run Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/s"

airspresso — Mon, 05 Oct 2026 05:45:40 +0000

> it's extremely long in it's thinking, it just goes on and on

Known issue with this model, I recommend setting thinking to 'medium' instead of default 'xhigh'.

New comment by airspresso in "Livenerf: Has Opus 5.5 been nerfed yet?"

airspresso — Wed, 30 Sep 2026 09:47:15 +0000

> I looked into running a local model, and it's way outside the budget of an individual or even a small or medium business (hundreds of thousands of dollars).

That is a big exaggeration. You can have a perfectly usable local LLM setup that will power your agent for single digit thousands of dollars. Can even power multiple agents simultaneously, depending on the hardware and setup. Won't be fast and won't be frontier intelligence, but definitely useful.

New comment by airspresso in "Early rogue AI agent activity and attempts to hack found on urlquery.net"

airspresso — Thu, 24 Sep 2026 09:41:04 +0000

> What if all the agents involved in these incidents have in fact had the full stack of alignment applied

A big part of this developing story is that it happened during training of a new model that ended up misaligned. And training happened without the usual safeguards applied like chain-of-thought monitoring. So OpenAI has already admitted that the full stack of aligment had certainly not been applied in this case.

New comment by airspresso in "Samsung is expected to more than double output of its HBM4 and HBM4E DRAM"

airspresso — Sun, 20 Sep 2026 19:16:51 +0000

We are certainly not in the diminishing returns phase for LLM progress. No sign of that yet.

New comment by airspresso in "Samsung is expected to more than double output of its HBM4 and HBM4E DRAM"

airspresso — Sun, 20 Sep 2026 19:14:57 +0000

No, those use LPDDR5(x), not HBM.

New comment by airspresso in "Notes on gotchas while migrating 35kb preprompts from Opus to self-hosted Ollama"

airspresso — Tue, 15 Sep 2026 06:07:47 +0000

From what I'm seeing elsewhere, context size up to 128k should be possible on this hardware. It really matters for agentic workloads to push that context size headroom up. Anthropic are spoiling us with models that do 500k context and beyond.

New comment by airspresso in "Making the Internet Boring"

airspresso — Thu, 03 Sep 2026 04:51:53 +0000

I'm with you on this, the knowledge sharing on YT is massive and in general not replicated in written form anywhere else. Anything that needs fixing involves a YT search for me these days, since that has the highest chance of turning up something useful and relevant. Only downside is that it's time consuming to find the right spot in the video and pause/rewind to understand how to do the thing properly. But having someone actually demonstrate how to do it, rather than just describing it with words, is so helpful and makes replicating it a lot easier.

New comment by airspresso in "Making the Internet Boring"

airspresso — Thu, 03 Sep 2026 04:41:44 +0000

Clicked the links just to see preview screenshots, that is hilarious with all the cows XD

New comment by airspresso in "Matrox: Graphics for Professionals"

airspresso — Mon, 31 Aug 2026 06:02:46 +0000

It gave the card an aura of mystique (pun intended) and I remember really wanting it. I guess that means they succeeded in making a statement with it.

New comment by airspresso in "Nvidia agrees to acquire Hugging Face for $13B"

airspresso — Thu, 27 Aug 2026 08:09:53 +0000

Nvidia is also pushing for local inference IMHO. They want open models and competition in the model layer, not two big labs controlling all of it.

New comment by airspresso in "OpenAI Jalapeño: Better than Nvidia Blackwell"

airspresso — Tue, 25 Aug 2026 19:22:06 +0000

This depends heavily on what the use-case is. Yes, if it's a coder making software and having to read LLM output then writing style matters. If the LLM is used in an automated data processing pipeline with a capped level of complexity, entirely different aspects matter and LLMs become more interchangeable.

New comment by airspresso in "OpenAI Jalapeño: Better than Nvidia Blackwell"

airspresso — Tue, 25 Aug 2026 19:19:18 +0000

By leveraging the experience Broadcom has in this area. Still remains to be seen how that goes when they want to scale production.

New comment by airspresso in "IPFS Maintainers Winding Down"

airspresso — Mon, 24 Aug 2026 18:20:42 +0000

Most of the generative art is hosted on IPFS.

New comment by airspresso in "Anthropic's best AI model struggles to attract users as cheaper tools thrive"

airspresso — Mon, 24 Aug 2026 07:37:35 +0000

Wait, it's allowed again? Completely missed that. Been avoiding to use it and trying to find workarounds, not great.

New comment by airspresso in "Cerebras CS-4"

airspresso — Wed, 19 Aug 2026 04:59:36 +0000

The Cerebras hardware is not locked to specific models / model families. Taalas is the company that's etching models into their silicon, locking it to that model forever.

New comment by airspresso in "Process as a Proxy for Motivation"

airspresso — Thu, 13 Aug 2026 05:08:27 +0000

Agree. My team standardized on black in our Python CI/CD pipeline and precommit hooks and it was surprising how much time we reclaimed from not having to review coding style and discuss patterns. The compromise was to apply black and not think about it. And it worked out really well. This was back in the days when humans still wrote code XD

New comment by airspresso in "Getting 25 Gbps Thunderbolt Ethernet on My Mac Studio"

airspresso — Fri, 31 Jul 2026 21:37:10 +0000

My default assumption (with no knowledge about this particular case) is that all gear in YouTube videos is sponsored. Too easy to get jealous of all the fancy setups. But I don't want to become a YouTuber so that settles that.

New comment by airspresso in "OpenAI unveils its first custom chip, built by Broadcom"

airspresso — Wed, 24 Jun 2026 17:49:24 +0000

OpenAI's upcoming mega IPO