Hacker News: galsapir

New comment by galsapir in "Openrouter Fusion API"

galsapir — Mon, 15 Jun 2026 11:44:15 +0000

really interesting that its basically almost 80% claude opus..

New comment by galsapir in "Openrouter Fusion API"

galsapir — Mon, 15 Jun 2026 11:42:16 +0000

yeah its really counterintuitive i think; i.e, getting the right framework and structure for this to work probably isn't trivial, models really hate playing well together. i wonder how their version would fair in real world use.

A bitter lesson for medicine, or a benchmark problem?

galsapir — Sun, 14 Jun 2026 21:30:42 +0000

Article URL: https://sparsethought.com/2026/06/14/what-did-they-actually-measure/

Comments URL: https://news.ycombinator.com/item?id=48532976

Points: 2

# Comments: 0

New comment by galsapir in "I procrastinate by building tools to stop me from procrastinating: A sad story"

galsapir — Thu, 11 Jun 2026 11:53:20 +0000

i feel like i've had exactly the same thought in the past :-0 might even have written about it. feel your pain

Can LLMs Beat Classical Hyperparameter Optimization Algorithms?

galsapir — Tue, 09 Jun 2026 15:01:15 +0000

Article URL: https://arxiv.org/abs/2603.24647

Comments URL: https://news.ycombinator.com/item?id=48462062

Points: 120

# Comments: 20

Gemma 4 E4B as a primary local LLM (replaced Qwen)

galsapir — Sun, 07 Jun 2026 18:49:43 +0000

Article URL: https://digg.com/ai/bfr4bqhh

Comments URL: https://news.ycombinator.com/item?id=48437537

Points: 2

# Comments: 0

New comment by galsapir in "My Agent Skill for Test-Driven Development"

galsapir — Fri, 05 Jun 2026 20:46:51 +0000

sometimes I also feel it tries to optimise for "per line coverage" over more "real, complex use cases" type tests

New comment by galsapir in "Hacker News front page as a site"

galsapir — Mon, 25 May 2026 20:19:23 +0000

hey that's pretty cool. I think I still prefer "distill HN" cleanliness though. What made you create this.

PEEK: Give Your Agent an Orientation Cache (MIT CSAIL, Khattab group)

galsapir — Mon, 25 May 2026 12:21:01 +0000

Article URL: https://zhuohangu.github.io/blog-post-peek/

Comments URL: https://news.ycombinator.com/item?id=48266008

Points: 3

# Comments: 0

New comment by galsapir in "Codex-maxxing"

galsapir — Sun, 24 May 2026 08:48:49 +0000

axon discharge is brilliant. adopting.

Hyperagents (Meta Research)

galsapir — Sat, 23 May 2026 14:02:30 +0000

Article URL: https://arxiv.org/abs/2603.19461

Comments URL: https://news.ycombinator.com/item?id=48247793

Points: 2

# Comments: 0

New comment by galsapir in "The Unreasonable Effectiveness of HTML"

galsapir — Thu, 21 May 2026 07:03:42 +0000

oh sorry! didn't catch the one Thanks, I'll comment there

The Unreasonable Effectiveness of HTML

galsapir — Thu, 21 May 2026 05:45:24 +0000

Article URL: https://claude.com/blog/using-claude-code-the-unreasonable-effectiveness-of-html

Comments URL: https://news.ycombinator.com/item?id=48218368

Points: 3

# Comments: 2

New comment by galsapir in "Gaussian Splat of a Strawberry"

galsapir — Tue, 19 May 2026 11:07:33 +0000

From the link: "Shot from 90 perspectives, 88 focus stacked images each. Nikon Z8, full frame, f/7.1, exposure 1/160, ISO 100, Laowa 180mm macro lens, with LED light and bluescreen." Insane!

New comment by galsapir in "Anthropic acquires Stainless"

galsapir — Tue, 19 May 2026 11:05:31 +0000

I think the question he tried to raise was "is this needed? Aren't today's / tomorrow's models well-enough equipped to deal with just OPEN API?" (idk, just if I understand the question)

New comment by galsapir in "Illusions of understanding in the sciences"

galsapir — Sun, 17 May 2026 18:58:24 +0000

got me at "Most often scientists believe they understand more than they do, making their belief an illusion." but why is it still bothering me? 1. feels unfalsifiable in spirit 2. somewhat restates "all models are wrong, but some are useful" less cleanly 3. doesn't really offer like, what can we do as science people? tomorrow morning perspective

New comment by galsapir in "The Comparator in Clinical AI"

galsapir — Thu, 07 May 2026 07:44:35 +0000

author here. the part i'd actually like discussion on is the buried finding: physicians+GPT-4 didn't outperform GPT-4 alone on the management cases, and on the landmark cases the model alone beat the model+physician. the paper reports it and moves on. that's the 2026 question, and it's the one a Science-level platform could have been used to ask

The Comparator in Clinical AI

galsapir — Thu, 07 May 2026 07:44:35 +0000

Article URL: https://sparsethought.com/2026/05/03/science-paper/

Comments URL: https://news.ycombinator.com/item?id=48046593

Points: 2

# Comments: 1

New comment by galsapir in "Borges' cartographers and the tacit skill of reading LM output"

galsapir — Sun, 12 Apr 2026 05:41:06 +0000

Hey thanks! I do wonder that. I think that even if specifically for code smell the things would be subtler, for other forms of AI driven averageness (especially in areas where we can't RLVR the models to perfection) it might still be present. But yeah I wonder how those thoughts will age (and how we'll update our priors accordingly).

New comment by galsapir in "Borges' cartographers and the tacit skill of reading LM output"

galsapir — Sat, 11 Apr 2026 19:09:57 +0000

yeah I was really thinking about what the best "umbrella term" would be here. Since "LLM" is too widely used in a really specific context and "AI systems" felt niche I ended up with "LMs". Idk, up for debate..