Hacker News: awwaiid

New comment by awwaiid in "Push events into a running session with channels"

awwaiid — Fri, 20 Mar 2026 02:56:19 +0000

Yes -- this is getting very close to ClaudeClaw. Next they'll offer cloud hosting of persistent execution.

New comment by awwaiid in "Show HN: Duplicate 3 layers in a 24B LLM, logical deduction .22→.76. No training"

awwaiid — Thu, 19 Mar 2026 02:50:43 +0000

Maybe another idea, no idea if this is a thing, you could pick your block-of-layers size (say... 6) and then during training swap those around every now and then at random. Maybe that would force the common api between blocks, specializaton of the blocks, and then post training analyze what each block is doing (maybe by deleting it while running benchmarks).

New comment by awwaiid in "Separating the Wayland compositor and window manager"

awwaiid — Mon, 16 Mar 2026 17:18:33 +0000

This was not my experience. For example, it want until chromium 110 that you could use webrtc+pipewire without overriding settings, in 2023. So maybe in a strict sense you could do it with fiddling, I don't think you could install any Linux flavor and screen share over meet reliably.

https://groups.google.com/g/discuss-webrtc/c/fe567r-UUrA/m/8...

New comment by awwaiid in "Malus – Clean Room as a Service"

awwaiid — Fri, 13 Mar 2026 13:21:20 +0000

.... did you give them money? Brave!

New comment by awwaiid in "Tennessee grandmother jailed after AI face recognition error links her to fraud"

awwaiid — Fri, 13 Mar 2026 01:29:54 +0000

Yes -- I know at least 3 Orion Blasters.

New comment by awwaiid in "Microgpt"

awwaiid — Sun, 01 Mar 2026 04:05:38 +0000

Where is this 1000 lines of C coming from? This is python.

New comment by awwaiid in "The path to ubiquitous AI (17k tokens/sec)"

awwaiid — Sat, 21 Feb 2026 03:07:09 +0000

Generative TikTok for words

New comment by awwaiid in "US plans online portal to bypass content bans in Europe and elsewhere"

awwaiid — Fri, 20 Feb 2026 02:32:51 +0000

Yep! Maximally closed as much as possible under the law. They also shut down other programs which aim to sidestep propaganda (including US propaganda), though some of those are starting to come back. Radio Free Asia, for example, https://www.reuters.com/world/asia-pacific/radio-free-asia-s...

New comment by awwaiid in "SkillsBench: Benchmarking how well agent skills work across diverse tasks"

awwaiid — Tue, 17 Feb 2026 01:14:00 +0000

I've been building a skill to help run manual tests on an app. So I go through and interactively steer toward a useful validation of a particular PR, navigating specifics of the app and what I care about and what I don't. Then in the end I have it build a skill that would have skipped backtracking and retries and the steering I did.

Then I do it again from scratch; this time it takes less steering. I have it update the skill further.

I've been doing this on a few different tests and building a skill which is taking less and steering to do app-specific and team-specific manual testing faster and faster. The first times through it took longer than manually testing the feature. While I've only started doing this recently, it is now taking less time than I would take, and posting screenshots of the results and testing steps in the PR for dev review. Ongoing exploration!

New comment by awwaiid in "How we automated federal retirements"

awwaiid — Sun, 28 Dec 2025 12:53:52 +0000

That was built recently and then discarded by this administration to protect intuit TurboTax and similar services.

New comment by awwaiid in "How we automated federal retirements"

awwaiid — Sun, 28 Dec 2025 12:52:20 +0000

That is very likely the main part they didn't choose so much as utilize an existing system, flowing the existing infrastructure.

New comment by awwaiid in "How we automated federal retirements"

awwaiid — Sun, 28 Dec 2025 12:49:22 +0000

It makes sense that this administration would focus investment on making it maximally easy to leave government above all other potential automations.

New comment by awwaiid in "How we automated federal retirements"

awwaiid — Sun, 28 Dec 2025 12:46:47 +0000

It was downplayed at every other opportunity including the conclusion, emphasizing instead the lone team of two heros. A few shout outs here and there, but the theme was clear.

New comment by awwaiid in "Tell HN: Happy LLM Day"

awwaiid — Sun, 30 Nov 2025 20:00:09 +0000

Fine... I guess it could be happy ChatGPT day then? Putting a hosted version of it out on the internet and letting people use it is very clearly a very small step, it even surprised OpenAI that it caught so many people's attention. But it did. It is an interesting milestone that shows what happens when you connect people with a new technology, a new concept, and see what they do with it.

New comment by awwaiid in "The Department of War just shot the accountants and opted for speed"

awwaiid — Tue, 11 Nov 2025 21:14:26 +0000

I was very confused until I realized the author was Steve Blank not Steve Klabnik.

New comment by awwaiid in "Claude Code on the Web: free usage credits"

awwaiid — Tue, 04 Nov 2025 23:14:57 +0000

This ploy is already working for me ... I had tried it before and now I'm trying it harder.

New comment by awwaiid in "Twake Drive – An open-source alternative to Google Drive"

awwaiid — Fri, 24 Oct 2025 16:04:14 +0000

Almost, but not entirely, unlike birds

New comment by awwaiid in "ChatGPT Atlas"

awwaiid — Tue, 21 Oct 2025 19:00:30 +0000

I asked Atlas about this, and it indirectly pointed out that atlas://credits is a thing. Not linked to anywhere that I could find though.

New comment by awwaiid in "Asciinema CLI 3.0 rewritten in Rust, adds live streaming, upgrades file format"

awwaiid — Sun, 21 Sep 2025 11:58:37 +0000

Very minor and tangential request for future recordings - show the shell prompt and/or put the commentary in shell comments (prefix with "# "). A step closer to IRL.

Speaking of which, an interesting thing to contemplate is if it is worth automating what you did, or if making the videos happens rarely enough that you'd start from scratch with a new manus or other ai session.

New comment by awwaiid in "Are OpenAI and Anthropic losing money on inference?"

awwaiid — Fri, 29 Aug 2025 10:59:08 +0000

Afaik all the large providers flipped the default to contractually NOT train on your data. So no, training data context size is not a factor.