yewenjie · HackerTrans

yewenjie

17 時間前議論

If only

yewenjie

24 日前議論

German startups that used to be English-speaking are now asking for German. I don't understand what that solves.

yewenjie

25 日前議論

Finding a job in tech that would genuinely excite me.

yewenjie

先月議論

So Dynamic Workflows is their version of ChatGPT Pro?

yewenjie

2 か月前議論

Is this a company and not a research lab doing this? What's the economic imperative for funding this?

yewenjie

2 か月前議論

> Chess is a lot trickier than it looks. It has so many rules: castling, en passant, pawn promotion, pinning, the discovered check, and the deadlock case of stalemate.

Nit: Pinning and the discovered check are not really rules, but rather names of tactics.

yewenjie

3 か月前議論

Interestingly ChatGPT right now answered

> Bixonimania is not a real disease. It was deliberately invented by scientists as an experiment to test whether AI systems and researchers would spread false medical information. Here’s the simple explanation ...

yewenjie

3 か月前議論

I love Lichess more than anything, and I hope this brings a lot of donation to them that they can use independently, and that the Lichess brand does not get subsumed by Take Take Take and their corporate money.

yewenjie

3 か月前議論

Are you building something that is novel, truly challenging, and meaningful? And looking for someone who thrives at really hard problems?

In the era of agents, the only real skill is thinking for yourself. That said, some background: master's in pure math, did research in AI/ML before LLMs, scaled early stage startups from scratch, been an open-source maintainer, and have a lot of peculiar interests.

I want to work with colleagues with a high degree of situational awareness, a solid understanding of cognitive biases, and an open mind. We will not get along if you have a strict opinion on how I should work.

Most recently I built data infrastructure for manufacturing at a startup in Germany. Before that, spent 3 years building and scaling a security data platform (open source, from early stage).

Happy to work in founding/staff engineer roles, or anything intellectually stimulating enough.

  Location: Berlin, Germany

  Remote: Yes

  Willing to relocate: Maybe, depends on how ambitious it is.
 
  Technologies: Pretty much anything at this point, any stack, all stack. Previously worked in web, machine learning, crypto. Favorite tech: functional programming, Nix, Emacs, but I work with whatever the problem needs like Typescript and Python and whatnot.

  Résumé/CV: https://shayon.work/cv.pdf

  Email: [email protected]

yewenjie

4 か月前議論

I have a Fuji X-T5 camera which I haven't really used, and have almost no workflow for.

Could someone explain what this software does?

And I'd also appreciate any software and workflow people use.

Would prefer Linux software, but macOS is okay if the quality/ ease of use is too different.

yewenjie

5 か月前議論

Does anyone know if it is just synthetic psilocybin or a psilocybin-like molecule?

yewenjie

5 か月前議論

One week ago I was on the hunt for an open source model that can do diatization and I had to literally give up because I could not find any easy to use setup.

yewenjie

6 か月前議論

Wait didn't xeiaso use NixOS?

https://xeiaso.net/talks/asg-2023-nixos/

yewenjie

7 か月前議論

We're truly living in reality that is much, much stranger than fiction.

Well, at least there's one company at the forefront that is taking all the serious issues more seriously than the others.

yewenjie

9 か月前議論

Is there any anecdata about emotional well-being for people without any conditions these medications are usually prescribed?

yewenjie

9 か月前議論

Does it support speaker diarization?

yewenjie

10 か月前議論

Looking at the chart here, it seems like Sonnet 4 was already better than GPT-5-codex in the SWE verified benchmark.

However, my subjective personal experience was GPT-5-codex was far better at complex problems than Claude Code.

yewenjie

10 か月前議論

I would not be comfortable sending my bank info passwords and all sorts of other sensitive data that I input and see on my screen to Gemini. How much is the qualitative performance difference with a local model?

yewenjie

10 か月前議論

6+ years shipping full-stack products, doing ML since 2018 before the LLM hype, maintainer of 100+ NixOS packages because I believe software should actually work.

  Location: Berlin, Germany
  Remote: Yes
  Willing to relocate: London/UK, SF/NYC/US
  Technologies: TypeScript, Python, React, SvelteKit, FastAPI, PostgreSQL, 
                PyTorch, TensorFlow, Effect-TS, Nix, Emacs
  Résumé/CV: https://shayon.work/cv.pdf
  Email: [email protected]

Currently Senior SWE at an early-stage cybersecurity startup where I built the entire frontend platform from scratch and architect complex web applications with heavy real-time data processing across distributed, user-controlled nodes. Before LLMs were cool, I was doing deep theoretical ML research (Inverse Reinforcement Learning) with an MS in Math.

Why I'm different: I single-handedly shipped our main product frontend from scratch, then scaled the team by hiring 3+ engineers - the kind of 0-to-1 execution that early-stage startups need. I tackle ambiguous problems with unconventional solutions, build tools that don't exist yet, and code across the full stack. Deep into Nix/NixOS and Emacs, heavy open source contributor.

Looking for senior/staff roles with real user impact, significant technical ownership, and complex problems to solve. Want passionate colleagues who love their craft and maintain their dotfiles with pride.

yewenjie

10 か月前議論

I wish some radical team just says fuck yes, we're gonna make Emacs fast and actually accomplishes it.

It's definitely easier with LLMs now, but still considerably hard.

コメント