Hacker News: dfsegoat

New comment by dfsegoat in "Kimi K2.5 Technical Report [pdf]"

dfsegoat — Sat, 31 Jan 2026 00:22:26 +0000

> but google is stock in the academic bias of Stanford and MIT and can't think outside the box

Can you clarify what you mean? I am not sure I follow.

New comment by dfsegoat in "Scaling LLMs to Larger Codebases"

dfsegoat — Mon, 22 Dec 2025 18:07:22 +0000

Highly recommend using agent based hooks for things like `[review & test]`.

At a basic level, they work akin to git-hooks, but they fire up a whole new context whenever certain events trigger (E.g. another agent finishes implementing changes) - and that hook instance is independent of the implementation context (which is great, as for the review case it is a semi-independent reviewer).

New comment by dfsegoat in "Beginning January 2026, all ACM publications will be made open access"

dfsegoat — Thu, 18 Dec 2025 17:36:06 +0000

It seems that perhaps neither are inherently 'good models'? What would an ideal alternative look like?

New comment by dfsegoat in "Gemini 3 Flash: Frontier intelligence built for speed"

dfsegoat — Wed, 17 Dec 2025 22:22:21 +0000

> it’s more performant than Claude Opus 4.5 or GPT 5.2 extra high

...and all of that done without any GPUs as far as i know! [1]

[1] - https://www.uncoveralpha.com/p/the-chip-made-for-the-ai-infe...

(tldr: afaik Google trained Gemini 3 entirely on tensor processing units - TPUs)

New comment by dfsegoat in "GPT-5.2"

dfsegoat — Fri, 12 Dec 2025 12:01:24 +0000

they 100% are unless you provide a RUBRIC / basically make it ordinal.

"Return a score of 0.0 if ...., Return a score of 0.5 if .... , Return a score of 1.0 if ..."

New comment by dfsegoat in "Claude outage"

dfsegoat — Fri, 31 Oct 2025 18:46:22 +0000

Very similar usage duration / window for me (US Pacific) - no availability issues.

New comment by dfsegoat in "Moderna has unraveled"

dfsegoat — Thu, 30 Oct 2025 18:12:25 +0000

To put some numbers to trying to develop a single therapy (where candidates etc. will fail as you try them)

- Plan to sink $180-500M+ just in R&D

- Factor in failures, regulatory, clinical, recruitment, phase 1/2 trials and you arrive very quickly around $1.3-2.1 BILLION USD per therapy approved.

...there is a 90% chance that you will spend that $1B+ - and it will fail completely.

https://www.nature.com/articles/d41573-020-00043-x

https://greenfieldchemical.com/2023/08/10/the-staggering-cos...

New comment by dfsegoat in "Israel demanded Google and Amazon use secret 'wink' to sidestep legal orders"

dfsegoat — Thu, 30 Oct 2025 17:53:54 +0000

fwiw towards your theory, I believe that the US Govt actually considers cloud providers - by way of specific services offered "dual use" systems for mil or civil use.

E.g. you will find references in AWS docs to Bureau of Industry/Security rulings.

https://en.wikipedia.org/wiki/Dual-use_technology

https://www.bis.gov/

https://aws.amazon.com/compliance/global-export-compliance/

New comment by dfsegoat in "Launch HN: Propolis (YC X25) – Browser agents that QA your web app autonomously"

dfsegoat — Thu, 30 Oct 2025 17:43:40 +0000

OTP protected flow verification

New comment by dfsegoat in "Launch HN: Propolis (YC X25) – Browser agents that QA your web app autonomously"

dfsegoat — Thu, 30 Oct 2025 17:43:01 +0000

Really good call out re: email and other 'side-flows' - hopefully there is integration with something like Mailosaur.

https://mailosaur.com/email-testing

New comment by dfsegoat in "The next chapter of the Microsoft–OpenAI partnership"

dfsegoat — Tue, 28 Oct 2025 16:04:16 +0000

> nor are they...able to mutate themselves to gain new information on the fly

See "Self-Adapting Language Models" from a group out of MIT recently which really gets at exactly that.

https://jyopari.github.io/posts/seal

New comment by dfsegoat in "OpenAI Is Just Another Boring, Desperate AI Startup"

dfsegoat — Fri, 03 Oct 2025 18:15:27 +0000

...not monetized yet: Can't find the post, but a prev. HN post had a link to an article showing that OpenAI had hired someone from Meta's ad service leadership - so I took that to mean it's a matter of time.

edit: believe it was Fidji Simo et al.

https://www.pymnts.com/artificial-intelligence-2/2025/openai...

New comment by dfsegoat in "Markov chains are the original language models"

dfsegoat — Thu, 25 Sep 2025 21:13:34 +0000

Ahh thank you for catching - apologies I didn't see this in time to add a correction.

New comment by dfsegoat in "Markov chains are the original language models"

dfsegoat — Wed, 24 Sep 2025 06:42:54 +0000

KingJamesProgramming [1] - a mashup of Knuth books [2] and the King James Bible with a Markov chain, is still one of my favorite reads for a laugh.

It was also the first place I was really exposed to probabilistic-based methods for generation of novel content.

You get gems like:

"they smote the city with the edge of the sword. 22:20 And one of his main motivations was the high cost of proprietary software houses."

and

"37:29 The righteous shall inherit the land, and leave it for an inheritance unto the children of Gad according to the number of steps that is linear in b."

[1] https://www.tumblr.com/kingjamesprogramming

[2] https://en.wikipedia.org/wiki/Donald_Knuth

New comment by dfsegoat in "A PM's Guide to AI Agent Architecture"

dfsegoat — Thu, 04 Sep 2025 18:57:58 +0000

I'm typically pretty critical of PM oriented pieces, but I found this to be a decent overview of how to reason about building these systems from first principles + some of the non-tech pain points + how to address them.

New comment by dfsegoat in "We're Joining OpenAI"

dfsegoat — Thu, 04 Sep 2025 18:43:42 +0000

Very good catch. For all my interest this space, it's embarrassing that I did not remotely see this coming...

New comment by dfsegoat in "We're Joining OpenAI"

dfsegoat — Wed, 03 Sep 2025 21:10:26 +0000

> Expect to see a lot of ads 2 years from now

I think I am just slow today - but could you please elaborate?

New comment by dfsegoat in "Intel's "Clearwater Forest" Xeon 7 E-Core CPU Will Be a Beast"

dfsegoat — Fri, 29 Aug 2025 15:55:16 +0000

Very curious about this as well.

Though I would say the fact that they were not involved with OpenAI on gpt-oss is not a good sign.

(OAI partnered with Databricks, NVIDIA, Dell, bunch of MSPs. etc.)

New comment by dfsegoat in "Hundreds lose water source in Colorado's poorest county with no notice"

dfsegoat — Mon, 25 Aug 2025 15:57:06 +0000

Colorado Water access law is truly crazy and arcane in some cases (such as this, it would seem).

Source: Had a friend in college that interned for a group of attorneys in Western, CO whose entire practice was around water access rights.

She explained to me some of the ridiculous things that neighbors requiring common water access could do to each other - based purely on who was using the water first.

New comment by dfsegoat in "Anthropic tightens usage limits for Claude Code without telling users"

dfsegoat — Thu, 17 Jul 2025 22:45:42 +0000

Can you clarify which tooling you are using? Is it cursor-stats?