News for hackers. Stuff that matters.

klntsky

Re:DeepSeek-v4.1-Flash-1017 on Saturday October 17, 2026 @11:03AM (#50131760)
Here's a small benchmark that compares the new checkout with the previous one:

https://slopgrid.com/deepseek-41-horses

Re:Jev Meets the Trolley Problem on Friday October 16, 2026 @10:07AM (#50118585)
Check out https://trolley.typememetics.institute/ - it's VERY related

Show HN: SlopGrid – Create your own oneshot AI model benchmarks

https://slopgrid.com/

This is something I've been playing with for the last couple days - an app I vibecoded to arrange oneshot LLM visual benchmarks without the need to write any code. I do it form my phone so the excellence of mobile UI is high in the priority list.It is inspired by Simon Willison's benchmark writeups[0], basically you can recreate them in one go.Besides text and images, it supports SVG outputs...

( Read More... | no comments | on HN )
tl;dr

> no mobile clients

It's unclear whether the reviews are just numerical noise or not.
What is the incentive for me to spend my tokens on submitting reviews?
True. I understand the author on the emotional level, but the fact that a lot of work went into the product does not mean that it is that valuable. The author should focus on making the product better - unfortunately (for them) that means inventing a new business model.
Re:Ask HN: Who wants to be hired? (October 2026) on Thursday October 01, 2026 @02:26PM (#49925270)

  Location: Georgia
  Remote: Yes
  Willing to relocate: Yes
  Technologies: TypeScript, LLM workflows, Rust, Haskell, Python (ML), web stack, linux, web3 (EVM). Always excited to learn something new
  Résumé/CV: https://klntsky.dev/klntsky-cv.pdf
  Email: my username at google mail
I build systems end to end, can lead your team or help with achieving a specific short term objective. I am good at translating requirements into architecture & avoiding unnecessary costs early. I can also teach you stuff (for example: FP, type-level TypeScript, how to prompt coding agents to achieve good outcomes)
It's strange that they did not mention reworking the tests themselves. In my experience AI-generated tests are absolute garbage - unless you specifically prompt to reason about coverage to avoid duplication and to merge some.

Show HN: Seamless pane navigation across nested tmux sessions over SSH

https://github.com/klntsky/tmux-seamless-remote

Hey, I've got tired of dealing with tmux prefixes so I made a plugin that makes pane switching seamless across SSH/nested tmux boundaries.Now you can just use Alt+↑/↓/←/→ instead of doing mental math calculating the right combination of prefixes to switch panes. Enjoy!

( Read More... | no comments | on HN )
There is a bit more mature project that does the same: https://selfprivacy.org
Re:Nitter has more working instances than before the takedowns on Saturday September 05, 2026 @02:53AM (#49573814)
Is there something like a balancer that redirects me to an instance that is not down?
Re:Record-High 89% in U.S. Say Government Corruption Widespread on Saturday September 05, 2026 @02:51AM (#49573809)
My impression is that americans got only those types of corruption most people can't benefit from personally (like skipping fines, getting some minor government benefits, local permits, getting legal work processed faster, etc). For most of them it's just on the TV and on the social media.
M-x doctor
Re:I set a trap for a book-marketing scammer (2025) on Sunday August 23, 2026 @01:08AM (#49406197)
It's sad that there are people who are creative enough to pull this off who choose to participate in petty scams instead of applying their ai and persuasion skills to something more useful.
Re:Palomar: A registry of Lean verified mathematics on Wednesday August 19, 2026 @06:29AM (#49359558)
No thats not a strange decision. self-hosted git servers disappear all the time, it's better to have everything unavailable at once when github is down than suffer whenever either of the sources goes unreachable. Non-developers view git availability as a simple utility, they don't attach a value judgement to git being usable with any remote.
What's the problem with connecting two devices together?
Re:How we used to get jobs: A newspaper classifieds story on Wednesday August 12, 2026 @02:24AM (#49268514)
Someone who flips cars serves the seller by closing the deal faster
Re:Stealing Reasoning Traces from Proprietary LLM APIs on Wednesday August 12, 2026 @12:38AM (#49267859)
The model is finetuned to enter/leave its thinking mode using special token separators. there's no reason to assume the tool calls induce the same token distribution or produce the model's actual native reasoning trace
Re:Ask HN: Who wants to be hired? (August 2026) on Tuesday August 04, 2026 @04:55AM (#49165960)

  Location: Georgia
  Remote: Yes
  Willing to relocate: Yes
  Technologies: TypeScript, LLM agents, Rust, Haskell, Python (ML), web stack, linux, web3 (EVM). Always excited to learn something new
  Résumé/CV: https://klntsky.dev/klntsky-cv.pdf
  Email: my username at google mail
I build systems end to end, can lead your team or help with achieving a specific short term objective. I am good at translating requirements into architecture & avoiding unnecessary costs. I can also teach you stuff (FP, type-level TypeScript, how to prompt coding agents to achieve good outcomes)
Readability is far from ideal. check this out: https://github.com/promptware/readweb

It's a custom pipeline I made to address its shortcomings

That is fair for something that makes the game possible in the first place.
It's insane that no one tried this internally during development
Re:Dependencies should be fetched directly from VCS on Sunday July 05, 2026 @06:12PM (#48798411)
The almost perfect solution is Nix
Re:The Log Is the Agent on Sunday July 05, 2026 @05:20AM (#48792597)
Can someone explain why such a trivial knowhow is paper-worthy? Event sourcing is well known
Re:Better Models: Worse Tools on Sunday July 05, 2026 @05:15AM (#48792569)
My assumption was that it is often not convenient if you have a lot of logic. I used it internally, and the complexity of my use cases was barely enough to justify it too. But I've seen systems where it would definitely be a value unlock if I had to integrate LLM chatbots into them
Re:Better Models: Worse Tools on Sunday July 05, 2026 @12:49AM (#48791337)
I've built a library that makes creating rich feedback systems easier, check this out:

https://tool2agent.org/

Re:Ask HN: Who wants to be hired? (July 2026) on Saturday July 04, 2026 @08:53AM (#48785080)

  Location: Georgia
  Remote: Yes
  Willing to relocate: Yes
  Technologies: TypeScript, LLM agents, Rust, Haskell, Python (ML), web stack, linux, web3 (EVM). Always excited to learn something new
  Résumé/CV: https://klntsky.dev/klntsky-cv.pdf
  Email: my username at google mail
I build systems end to end, can lead your team or help with achieving a specific short term objective. I am good at translating requirements into architecture, avoiding unnecessary costs, and helping people discover their strengths under supervision. I can also teach you stuff (FP, type-level TypeScript, how to prompt coding agents to achieve good outcomes)
Re:Claude Code Is Steganographically Marking Requests on Tuesday June 30, 2026 @03:13PM (#48737842)
I would add that it would probably work even better than a KYC at least for some time until discovered, given that there is a very developed international market for KYC bypass services

Older »