Hacker News: mohsen1

Issues Reaching Status.claude.com

mohsen1 — Fri, 14 Aug 2026 08:17:00 +0000

Article URL: https://anthropic.statuspage.io/incidents/kmbpgrsszf72

Comments URL: https://news.ycombinator.com/item?id=49295947

Points: 3

# Comments: 0

New comment by mohsen1 in ""The Persian MâR-Nâmeh Or, the Book for Taking Omens from Snakes" (1892)"

mohsen1 — Mon, 10 Aug 2026 08:57:34 +0000

wow i totally missed "It Was Just an Accident". Have to watch asap.

New comment by mohsen1 in "Interviewing Engineers in the AI Era: Lessons from a Year of Rebuilding"

mohsen1 — Fri, 07 Aug 2026 11:33:30 +0000

Hmmm... I had a different experience. They had a fully automated environment where you ha to write test that passes some tests. No human involved. And the time requirement was insanely tight

New comment by mohsen1 in "When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation"

mohsen1 — Tue, 04 Aug 2026 22:01:09 +0000

That's a lot of words to say you need larger set of questions for today's models. 300 questions won't be enough to find the difference

New comment by mohsen1 in "Kimi Linear: An Expressive, Efficient Attention Architecture"

mohsen1 — Tue, 28 Jul 2026 15:57:56 +0000

> one of the basic tenants of algorithm development was that you can't just brute-force your way towards a solution for some complex problems

Mote-Carlo is pretty useful still. Not sure if your statement holds

New comment by mohsen1 in "Kimi K3 exploited the latest Redis server"

mohsen1 — Fri, 24 Jul 2026 22:54:19 +0000

Do you think any programmer really understands how their program works end-to-end? At some abstraction layer, we're all clueless. There are many layers between what you type into the text editor and the actual CPU ticks that make your program work. I bet nobody fully understands the whole stack.

Now that that text editor accepts English, we're all calling each other names, etc.

New comment by mohsen1 in "Claude Opus 5"

mohsen1 — Fri, 24 Jul 2026 21:50:19 +0000

RL. Lots of RL

New comment by mohsen1 in "Sleep regularity is a stronger predictor of mortality risk than sleep duration (2023)"

mohsen1 — Wed, 15 Jul 2026 13:50:26 +0000

I thought the same. If your life is so in order that you routinely sleep on the same interval, perhaps your life is not as stressful as others who sleep more chaotically

New comment by mohsen1 in "Ask HN: What Are You Working On? (July 2026)"

mohsen1 — Mon, 13 Jul 2026 07:15:52 +0000

This is a Chrome extension that records lots of details in a usage session. Stuff like network calls, console logs, screenshots and also optionally screenshots and user narration

Tools like this exist, but every one I tried is uploading the session details somewhere in their cloud and try to monetize this.

So I built the version I wanted: free, open source, and local. There is no account, no backend, no telemetry. Sessions live in IndexedDB in your browser and exported as a zip.

What it records:

* Clicks, typing, page changes, network requests and responses, console errors screenshots, video with sound

* Your voice, transcribed and placed next to what you were doing at the time

* Annotations: Arrows and boxes you draw on the page's screenshot

Note: Passwords, auth headers, and tokens are masked at capture time

All events are lined up in a timeline with timestamps

At export you pick a detail level with a live token estimate, so a long session still fits your model's context window.

Repo: https://github.com/mohsen1/session-recorder-chrome-extension

Show HN: A free and open source session recording Chrome extension

mohsen1 — Fri, 10 Jul 2026 23:05:30 +0000

Hey HN,

This is a Chrome extension that records lots of details in a usage session. Stuff like network calls, console logs, screenshots and also optionally screenshots and user narration

Tools like this exist, but every one I tried is uploading the session details somewhere in their cloud and try to monetize this.

So I built the version I wanted: free, open source, and local. There is no account, no backend, no telemetry. Sessions live in IndexedDB in your browser and exported as a zip.

What it records:

* Clicks, typing, page changes, network requests and responses, console errors screenshots, video with sound

* Your voice, transcribed and placed next to what you were doing at the time

* Annotations: Arrows and boxes you draw on the page's screenshot

Note: Passwords, auth headers, and tokens are masked at capture time

All events are lined up in a timeline with timestamps

At export you pick a detail level with a live token estimate, so a long session still fits your model's context window.

Chrome Web Store: https://chromewebstore.google.com/detail/session-recorder/db...

Repo: https://github.com/mohsen1/session-recorder-chrome-extension

Comments URL: https://news.ycombinator.com/item?id=48866471

Points: 3

# Comments: 0

New comment by mohsen1 in "Cargo-nextest: 3x faster than cargo test, per-test isolation, first-class CI"

mohsen1 — Thu, 09 Jul 2026 09:19:41 +0000

Thanks! Any pro tips for sharding? I landed on single job because couldn't get cache to work properly for shards to be fast enough to worth it

New comment by mohsen1 in "Cargo-nextest: 3x faster than cargo test, per-test isolation, first-class CI"

mohsen1 — Thu, 09 Jul 2026 08:25:17 +0000

I love nextest. without it my CI could take hours

https://github.com/tsz-org/tsz/actions/runs/29002057457/job/...

watch it running 32.5k unit tests without breaking a sweat!

New comment by mohsen1 in "Cloudflare Drop"

mohsen1 — Thu, 09 Jul 2026 07:42:23 +0000

This is perfect for my Chrome Extension for recording sessions and capturing screenshots, audio narration and videos. The output is a zip file with everything so if user wants to share they can use this

https://github.com/mohsen1/session-recorder-chrome-extension

I built above chrome extension because anything in this area has been trying to monetize the solution. I wanted a free and open source version of this to exist.

New comment by mohsen1 in "GPT-5.5 Codex reasoning-token clustering may be leading to degraded performance"

mohsen1 — Sun, 05 Jul 2026 08:21:48 +0000

I have a philosophical problem with adaptive thinking. It’s a dumb guess for how much thinking budget to allocate ahead of thinking. At least in the context of LLMs there is probably no way of knowing how much thinking (token generation) is needed. The problem space is infinity vast, similarly of two prompts is not going to help any LLM decide how much thinning is needed. Models already stop thinking before hitting the thinking budget.

Why there is so much effort in making adaptive thinking happen and don’t we train models to produce the end of thinning token better?

Feels like a bandaid. We need models to be trained to do a reasonable amount of reasoning (no pub intended):

    reason

    estimate remaining uncertainty

    continue?

    reason more

    repeat

New comment by mohsen1 in "Opening up 'Zero-Knowledge Proof' technology to promote privacy in age assurance"

mohsen1 — Thu, 02 Jul 2026 14:30:56 +0000

Computer Scientist Explains One Concept in 5 Levels of Difficult

https://www.youtube.com/watch?v=fOGdb1CTu5c

This video is very beautifully explains it

New comment by mohsen1 in "Micro-Agent: Beat Frontier Models with Collaboration Inside Model API"

mohsen1 — Mon, 29 Jun 2026 23:49:15 +0000

This seems to be a new trend. Noticed it with GPT "ultra" in their announcement[1]. I'm with you, a large language model and a system of many language models working together are not the same thing

[1] https://news.ycombinator.com/item?id=48689338

New comment by mohsen1 in "Do LLMs pass the mirror test?"

mohsen1 — Sun, 28 Jun 2026 22:27:13 +0000

It seems like we forget that LLMs are next token prediction systems. Using raw models without instruction following and chat completion bells and whistles will give you a better feeling of what LLMs are.

The current interface to LLMs are heavily biased towards "predict the next token in the context of a user with a helpful assistant" but LLMs are capable of other modes of next token prediction too.

Before the ChatGPT release people often measured LLM performance by how well they could produce a coherent story or a poem. that's where Anthropic model names are originating from I am guessing.

New comment by mohsen1 in "MicroVMs: Run isolated sandboxes with full lifecycle control"

mohsen1 — Fri, 26 Jun 2026 21:37:42 +0000

Nice thing about the microvm is that you can snapshot it and restore it. Keeping the fs minimal is my goal. Snapshots restore much faster if they do not include 5GB of source code

New comment by mohsen1 in "MicroVMs: Run isolated sandboxes with full lifecycle control"

mohsen1 — Fri, 26 Jun 2026 19:46:28 +0000

I’ve been working with AgentCore that uses the same MicroVMs. They are capable in many ways but for coding agents that load a big got repo they get bloated quickly with the git repo.

I’m building this google3 style mounting to address this.

https://github.com/mohsen1/git-lazy-mount

Still work in progress but for now I am seeing promising results

New comment by mohsen1 in "Previewing GPT‑5.6 Sol: a next-generation model"

mohsen1 — Fri, 26 Jun 2026 18:04:26 +0000

Confusion is: ultracode is not a different model with its own benchmarks