Hacker News: richardfey

New comment by richardfey in "Total Annihilation, rebuilt for modern computers"

richardfey — Mon, 12 Oct 2026 07:44:07 +0000

Tried to use it and the error messages and documentation were overwhelmingly unhelpful. Spent about an hour on it. Maybe it will get better though, good luck!

I wish the useful improvements would be ported to https://github.com/open-annihilation/open-annihilation which seems a better open source project.

New comment by richardfey in "Anthropic appears to be A/B testing reduced effort levels in Claude Code"

richardfey — Sun, 23 Aug 2026 04:21:42 +0000

Yes; or something which has a similar effect.

New comment by richardfey in "Anthropic appears to be A/B testing reduced effort levels in Claude Code"

richardfey — Sat, 22 Aug 2026 22:57:06 +0000

It feels lke they replace older models with "optimised" versions which are cheaper to run, but keep the same name.

New comment by richardfey in "How do I permanently disable random Google Photos popup to backup photos? (2024)"

richardfey — Mon, 17 Aug 2026 05:20:49 +0000

For pixel owners: the Google photos app cannot be uninstalled, but you can disable it using adb.

New comment by richardfey in "Red queen hypothesis – A new way forward for self-improving AI"

richardfey — Mon, 17 Aug 2026 05:13:43 +0000

> "Instead of improving an agent against a fixed test, we let the evaluation evolve alongside the agent"

This quote should have been highlighted earlier in the article.

New comment by richardfey in "Why does Opus 5 feel worse to work with?"

richardfey — Fri, 14 Aug 2026 19:39:45 +0000

It defines and introduces a lot of concepts/acronyms in the thinking blocks which we normally don't read.

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

richardfey — Wed, 12 Aug 2026 20:46:33 +0000

> If the model was designed specifically to quantize down to 1.58b, then it's different.

> AFAIK, there's no large models designed for this yet.

Isn't BitNet b1.58 2B4T what you are looking for? (haven't tried it myself though)

New comment by richardfey in "New Bedford police officer accused of using Flock cameras to track ex-partner"

richardfey — Wed, 12 Aug 2026 05:58:08 +0000

> Quis custodiet ipsos custodes? is a question that is roughly 2000 years old. Still no good answer.

I feel like we are moving towards a very specific answer, without much thought: AI.

New comment by richardfey in "Claude Code pricing: same tokens, same model, up to 40x the price"

richardfey — Tue, 11 Aug 2026 11:33:12 +0000

Did anyone calculate this for OpenAI?

New comment by richardfey in "Meta Muse Glimmer – open weights 30B local coding model"

richardfey — Mon, 10 Aug 2026 10:51:18 +0000

Looking forward to giving this a try with llama.cpp. I’m watching the open-weights competition with high expectations.

New comment by richardfey in "Lost my phone at the office. Claude suggested tracking Bluetooth signal strength"

richardfey — Sat, 08 Aug 2026 21:10:00 +0000

I think that would spoil a lot of the fun. I would favour visual and audio feedback loops to interactively learn how things work under the hood.

New comment by richardfey in "Lost my phone at the office. Claude suggested tracking Bluetooth signal strength"

richardfey — Sat, 08 Aug 2026 21:03:19 +0000

Sounds like a great way to inflate your success metrics for AI queries

New comment by richardfey in "Ten advances in mathematics and theoretical computer science"

richardfey — Thu, 06 Aug 2026 17:38:06 +0000

I disagree. Whether something is manipulation depends on whether you are trying to change someone’s opinion or behavior, not on whether the manipulator has “good” or “bad” intentions, since those judgments are not objectively universal.

New comment by richardfey in "Ten advances in mathematics and theoretical computer science"

richardfey — Tue, 04 Aug 2026 06:30:17 +0000

> Part of this can be good (you talk about what they care about, where 90% of broadcast messaging might not apply) and part of it can be bad (manipulation.)

Side note: it's manipulation either ways because you chose what to talk about, with a goal in mind.

New comment by richardfey in "Harness Engineering"

richardfey — Sun, 19 Jul 2026 09:25:55 +0000

Serious question: how do we verify claims like these on the effectiveness of a harness?

New comment by richardfey in "The Kimi K3 Moment"

richardfey — Sat, 18 Jul 2026 20:41:33 +0000

Has anyone tried Kimi K3 against gpt-5.6-sol on real projects?

New comment by richardfey in "The Kimi K3 Moment"

richardfey — Sat, 18 Jul 2026 20:33:53 +0000

Exactly; this is a no-go for me, I will wait for an independent provider to sell the service, which is possible thanks to the open weights.

New comment by richardfey in "Kimi K3: Open Frontier Intelligence"

richardfey — Fri, 17 Jul 2026 20:38:08 +0000

I understand you're trying to be funny, but my point is that with novel technology there are novel ways to claim innocence in courts because of the legislative void.

New comment by richardfey in "Kimi K3: Open Frontier Intelligence"

richardfey — Thu, 16 Jul 2026 22:38:49 +0000

They could use an agent to summarise the source material, and then train models on those summaries, and claim that some sort of clean-room training has happened?

New comment by richardfey in "Meta's Un-Stable Signature"

richardfey — Tue, 07 Jul 2026 12:08:15 +0000

> Or are you referring to the claims from Meta, Google, and Adobe -- which failed to hold up under independent evaluation.

This.

> However, he demonstrated a clear lack of understanding regarding what makes the bits "independent" or how to resolve the independence problem.

Yes, I didn’t want to call it out explicitly, but this is exactly the kind of thing that would have made my undergraduate statistics professor lose patience.