Hacker News: shay_ker

New comment by shay_ker in "The Siri for Families Apple Will Never Build"

shay_ker — Thu, 14 May 2026 12:57:43 +0000

hm yeah. it could be nice for apple to have a list of shortcuts that'd actually be useful based on real activity. but getting all the info needed is hard.

New comment by shay_ker in "Reverting the incremental GC in Python 3.14 and 3.15"

shay_ker — Thu, 14 May 2026 12:50:10 +0000

It's really fascinating to read this, since I've encountered similar memory issues in other languages (ruby, go, etc.). Debugging these issues is a pain.

Is there a way to make all this much easier to debug and to prevent memory issues in the first place? Is the abstraction level not quite right?

New comment by shay_ker in "DeepSeek 4 Flash local inference engine for Metal"

shay_ker — Thu, 07 May 2026 15:54:12 +0000

How does it compare to popular local inference engines, e.g. ollama, lm studio, or handrolled llama.cpp? I saw a brief benchmark in the readme but wasn't sure if there was more.

New comment by shay_ker in "Accelerating Gemma 4: faster inference with multi-token prediction drafters"

shay_ker — Tue, 05 May 2026 17:18:23 +0000

curious that they are doing speculative decoding and not baking MTP into the model, like Nemotron

https://docs.nvidia.com/megatron-core/developer-guide/0.15.0...

New comment by shay_ker in "Redis array: short story of a long development process"

shay_ker — Mon, 04 May 2026 21:37:49 +0000

antirez: i'm curious, with the final code, have you experimented with effectively one-shotting the final result? i wonder if we can get there with GEPA, and maybe there's something we can learn in how to elicit/prompt these models to get what we want.

or maybe the conclusion is that model providers need to clean up their training data!

Claude Design Is Real Design

shay_ker — Sun, 26 Apr 2026 21:06:37 +0000

Article URL: https://diverging.run/checkpoints/claude-design-is-real-design/

Comments URL: https://news.ycombinator.com/item?id=47914468

Points: 2

# Comments: 0

New comment by shay_ker in "Open Source Isn't Dead. Cal.com Just Learned the Wrong Lesson"

shay_ker — Wed, 15 Apr 2026 16:27:08 +0000

It's a good question - is blackbox hacking as effective as whitebox hacking, for AI agents? I've gotta assume someone at Anthropic is putting together an eval as we speak.

New comment by shay_ker in "Instant 1.0, a backend for AI-coded apps"

shay_ker — Thu, 09 Apr 2026 23:23:59 +0000

with a huge multi-tenant database, how do you deal with noisy neighbors? indexes are surely necessary, which impose a non-trivial cost at scale.

New comment by shay_ker in "Reinventing the Pull Request"

shay_ker — Thu, 02 Apr 2026 14:14:45 +0000

I'd love to see an example PR on lubeno vs github. That might show the comparison a bit better.

New comment by shay_ker in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

shay_ker — Tue, 24 Mar 2026 13:47:48 +0000

A general question - how do frontier AI companies handle scenarios like this in their training data? If they train their models naively, then training data injection seems very possible and could make models silently pwn people.

Do the labs label code versions with an associated CVE to label them as compromised (telling the model what NOT to do)? Do they do adversarial RL environments to teach what's good/bad? I'm very curious since it's inevitable some pwned code ends up as training data no matter what.

RocksDB unit test finds a CPU bug

shay_ker — Mon, 23 Mar 2026 14:36:19 +0000

Article URL: https://rocksdb.org/blog/2026/02/17/cpu-bug.html

Comments URL: https://news.ycombinator.com/item?id=47490149

Points: 3

# Comments: 0

New comment by shay_ker in "My spicy take on vibe coding for PMs"

shay_ker — Wed, 04 Mar 2026 03:48:56 +0000

I remember this post. But I'm not sure what the future really entails and I suspect it'll be very company/culture dependent. In some companies, the engineers are very savvy and understand the business well. In others, it's the designers. Or sales. Ops. And of course Product Managers. You get the picture.

Whoever gets the business best (and in detail) will likely be the best builders. It's "intuition as evals" that really matters in the end. You think Software Engineers or Product Managers are replacing Quants at trading shops anytime soon? Nope.

New comment by shay_ker in "430k-year-old well-preserved wooden tools are the oldest ever found"

shay_ker — Tue, 27 Jan 2026 19:43:12 +0000

The thing I’m continually surprised by is the usage of obsidian by nearly every ancient-ish civilization. The usage of bow & arrow predates farming, insane.

New comment by shay_ker in "Unison 1.0"

shay_ker — Tue, 25 Nov 2025 22:52:02 +0000

Very dumb question - sending code over the network to be executed elsewhere feels like a security risk to me?

I’m also curious how this looks with browser or mobile clients. Surely they’re not sending code to the server?

New comment by shay_ker in "Hello-World iOS App in Assembly"

shay_ker — Thu, 30 Oct 2025 16:15:55 +0000

This is the most HN comment I’ve seen in a while. The real abstraction here is coding with LLMs btw!!!

New comment by shay_ker in "A generic non-invasive neuromotor interface for human-computer interaction"

shay_ker — Thu, 24 Jul 2025 14:12:58 +0000

Associated GitHub repo: https://github.com/facebookresearch/generic-neuromotor-inter...

New comment by shay_ker in "WebRTC for the Curious"

shay_ker — Sat, 12 Apr 2025 15:49:13 +0000

Unfortunately, there's massive drift between WebRTC implementations. The status quo is Google's, which is used by Chrome: https://webrtc.googlesource.com/src/

And everyone else has to play catch-up.

A big source of the drift is having a common library for SDP parsing, but also necessary features like BWE, different encodings, etc.

For example, aiortc, python's WebRTC implementation, isn't quite at the level most would want. It isn't necessarily easy to tell, without clear benchmarks, which implementation is at-par with Google's.

India forces Namecheap to take down thejuggernaut.com (YC W19)

shay_ker — Wed, 08 Jan 2025 13:45:39 +0000

Article URL: https://twitter.com/namecheapceo/status/1875982043051430000

Comments URL: https://news.ycombinator.com/item?id=42634125

Points: 1

# Comments: 0

New comment by shay_ker in "Ask HN: Platform for senior devs to learn other programming languages?"

shay_ker — Sat, 30 Nov 2024 00:26:12 +0000

I'm a big fan of Execute Program from Gary Bernhardt:

https://www.executeprogram.com/

It uses interactive exercises + spaced repetition. It's really the only way I've managed to learn Regular Expressions.

New comment by shay_ker in "Hetzner Cloud Expands to Singapore"

shay_ker — Tue, 06 Aug 2024 13:36:42 +0000

Don’t AWS and GCP also produce some of their own chips and hardware?

But basically you’re saying, yes, AWS and GCP are making just as much, maybe more, margin as Hetzner?