Hacker News: gilgoomesh

New comment by gilgoomesh in "Felix Rieseberg Homepage"

gilgoomesh — Wed, 14 Oct 2026 06:37:06 +0000

It feels like it's trying to evoke Bioshock vibes or a similar retro-punk dystopia.

New comment by gilgoomesh in "Qwen 3.8 27B available on Cerebras at 1500 tokens/s"

gilgoomesh — Fri, 04 Sep 2026 08:57:53 +0000

It is common for the agent tools, MCP tool, skills, project context and system prompt to take between 5k and 25k tokens. This depends heavily on your agent and tool setup.

But that's before the prompt.

Then after the prompt, every tool call the agent makes adds to the context. Longer turns can easily consume 50k-100k tokens between the agent and various tool calls (reading the filesystem, reading files, reading compiler output, reading memories).

Then each "turn" with the agent stays in context and is fed into the next turn. Two or three turns and you're up near 250k.

New comment by gilgoomesh in "Qwen 3.8 27B available on Cerebras at 1500 tokens/s"

gilgoomesh — Fri, 04 Sep 2026 08:50:01 +0000

Easy. I just need... 50 times more speed.

Better download some more RAM.

New comment by gilgoomesh in "Pressed Penny Machine Map"

gilgoomesh — Wed, 19 Aug 2026 21:48:51 +0000

Destroying currency is crime in Australia, so no, this has never been a thing, here.

New comment by gilgoomesh in "Qwen3.8-2.4T"

gilgoomesh — Thu, 13 Aug 2026 07:45:42 +0000

Yeah, I must have misread the press release last week as I thought it would be released at the same time.

New comment by gilgoomesh in "Qwen3.8-Max: A New Bar for Coding and Cowork"

gilgoomesh — Mon, 03 Aug 2026 06:31:55 +0000

It's been 2.5 months since Qwen 3.7-Max. Unless you mean since the preview of this 3.8-Max model.

New comment by gilgoomesh in "Qwen3.8-Max: A New Bar for Coding and Cowork"

gilgoomesh — Mon, 03 Aug 2026 05:55:49 +0000

> lack of European competitors

There's Mistral. Although they've been surprisingly quiet for the last 6 months.

New comment by gilgoomesh in "Half-Life ported to Mac OS 9"

gilgoomesh — Wed, 29 Jul 2026 02:31:21 +0000

In one of the most infamous stories of the era, Half-Life was officially ported to Mac OS back in the year 2000 by Logicware/MacPlay but got shut down by Valve at the last minute.

I've watched interviews with Rebecca Heineman (one of the devs involved) on the topic. It did not make her happy.

Here's an article from the era:

https://www.gamespot.com/articles/mac-half-life-canceled/110...

Interview with Rebecca Heineman:

https://www.youtube.com/watch?v=lPTLPXNtb2I

(skip to timecode 1:36:41 for the Half-Life details)

New comment by gilgoomesh in "Claude Fable produced a counterexample to the Jacobian Conjecture"

gilgoomesh — Mon, 20 Jul 2026 10:29:22 +0000

...not going anywhere.

New comment by gilgoomesh in "Claude Fable produced a counterexample to the Jacobian Conjecture"

gilgoomesh — Mon, 20 Jul 2026 10:28:30 +0000

Goldbach's Conjecture is very accessible.

New comment by gilgoomesh in "An OpenAI model has disproved a central conjecture in discrete geometry"

gilgoomesh — Thu, 21 May 2026 05:31:12 +0000

I'm disappointed only that the chain of thought needed to be rewritten. Need to train these LLMs to natively communicate in LaTeX research paper format.

New comment by gilgoomesh in "Codex Hacked a Samsung TV"

gilgoomesh — Thu, 16 Apr 2026 23:11:20 +0000

> It’s not hacking, I wasn’t breaking anything

That's a very narrow read of the word "hacking".

We're literally on a website called "Hacker News". We're not all trying to break things.

New comment by gilgoomesh in "All elementary functions from a single binary operator"

gilgoomesh — Mon, 13 Apr 2026 04:30:20 +0000

> Why use splines or polynomials or haphazardly chosen basis functions if you can just fit (gradient descent) your data or wave functions to the proper computational EML tree?

Same reason all boolean logic isn't performed with combinations of NAND – it's computationally inefficient. Polynomials are (for their expressivity) very quick to compute.

New comment by gilgoomesh in "Sam Altman may control our future – can he be trusted?"

gilgoomesh — Tue, 07 Apr 2026 02:47:54 +0000

I think you can look at Star Trek as a fairly grounded example of where current LLMs could go: the ship's computer is not autonomous in any way but it does accept fairly vague instructions and you can apparently vibe-code the holodeck.

New comment by gilgoomesh in "Blood test boosts Alzheimer's diagnosis accuracy to 94.5%, clinical study shows"

gilgoomesh — Tue, 24 Feb 2026 04:51:38 +0000

For 20-ish% of Alzheimer's patients, the Shingles vaccine may be a treatment. This has been suspected for a few years now but has received recent confirmation studies.

https://www.alzheimers.org.uk/news/2025-11-18/promising-rese...

New comment by gilgoomesh in "Google Public CA is down"

gilgoomesh — Wed, 18 Feb 2026 02:05:45 +0000

It seems to be back, now.

New comment by gilgoomesh in "Launching the Handmade Software Foundation"

gilgoomesh — Sun, 18 Jan 2026 07:54:10 +0000

I read that but it doesn't define handmade. It gripes about large frameworks and rewriting in different languages but doesn't say what handmade is or how it addresses anything.

New comment by gilgoomesh in "Launching the Handmade Software Foundation"

gilgoomesh — Sun, 18 Jan 2026 07:51:44 +0000

Do they define "Handmade"? I couldn't find a definition.

New comment by gilgoomesh in "Apple Creator Studio"

gilgoomesh — Wed, 14 Jan 2026 00:00:53 +0000

You got Liquid Metal? I'm stuck with Liquid Glass :-(

New comment by gilgoomesh in "Apple picks Gemini to power Siri"

gilgoomesh — Mon, 12 Jan 2026 23:23:13 +0000

That 3B model is a local model that eventually got built into macOS 26. Gemini 3 Pro is a frontier model (cloud). They're very different things.