Hacker News: bfeynman

New comment by bfeynman in "Claude Fable 5"

bfeynman — Tue, 09 Jun 2026 17:59:25 +0000

Given it was made by cognition (team behind devin flop) who now just got to wait out until claude and gpt5 basically do all of the work for them - not very. When you read about it, the framework is highly subjective. Which very quickly becomes a problem because its based on heuristics that probably change a bunch with a better code model.

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

bfeynman — Sat, 06 Jun 2026 02:06:50 +0000

In what world or frame of reference would doing TDD have "little" bearing on output quality? If you build a system around satisfying some set of requirements it seems logical that output quality would have pretty heavy correlation.

New comment by bfeynman in "India's surprise baby bust"

bfeynman — Sat, 06 Jun 2026 01:51:21 +0000

I hate to break it to you but those are almost all economic problems in the grand scheme of things.

New comment by bfeynman in "India's surprise baby bust"

bfeynman — Sat, 06 Jun 2026 01:49:31 +0000

This is talked about all the time in the US as a huge impending problem with articles about all the time. You have segment of population with higher education and higher wealth that are having fewer children and later in life as it takes more time to get secured financially and also children are very expensive in terms of maintaining what used to be a middle class life style.

New comment by bfeynman in "Conventional Commits encourages focus on the wrong things"

bfeynman — Fri, 05 Jun 2026 22:39:26 +0000

why is "solving" the issue somehow the bar? software engineering has more practices rooted in psychology than engineering, its a moving and ambiguous target. Using conventional commits gives you a framework and mechanism that undoubtedly improves contribution semantics.

New comment by bfeynman in "Conventional Commits encourages focus on the wrong things"

bfeynman — Fri, 05 Jun 2026 22:34:36 +0000

From my own experience in big tech and various OSS things, the crux of this piece rests on this whole scope not being focal point misses that in general the other practice is making sure PRs have limited scope in general so it's marginalized, i.e. reducing blast radius or too big of a change at once by design is implied.

New comment by bfeynman in "New York to tax luxury second homes in NYC"

bfeynman — Tue, 19 May 2026 18:19:05 +0000

You are missing the point. Of course people want to live in one of the most vibrant cities in the world. People also want a vacation home at the beach or in the mountains thats private and beautiful and easy to get to. Except if we built giant monstrosities and condos in the hamptons and make all ski homes tenement housing it will be much less desirable to go to them. No ones asking to make more apartments and housing in rust belt cities.

New comment by bfeynman in "New York to tax luxury second homes in NYC"

bfeynman — Tue, 19 May 2026 18:14:14 +0000

Eh, I would only think if its like at the top of the park where less people/tourists ever visit.

New comment by bfeynman in "New York to tax luxury second homes in NYC"

bfeynman — Tue, 19 May 2026 18:10:54 +0000

I am a New Yorker, people want more housing but there is still NIMBYism because they want to preserve the charm, and I'm mostly only talking about manhattan. While people are not fans of the low density luxury skyscrapers popping up in places, I've not seen people who currently live in the place think we should add massive housing blocs carte blanche. Sure there few scattered places for a few projects but not like advocating to tear down to build bigger. That mentality comes from people who are definitely not new yorkers or live in fringes.

New comment by bfeynman in "Cursor Introduces Composer 2.5"

bfeynman — Tue, 19 May 2026 16:15:14 +0000

That only seems plausible if whatever corpse of xAI is around is giving them engineering time. I don't know if they hired a bunch of ex frontier lab staff but its unlikely they have the technical capability to train their own frontier models especially the pretraining. Because the thing is if its not competitive with claude/codex it will be panned.

New comment by bfeynman in "New York to tax luxury second homes in NYC"

bfeynman — Tue, 19 May 2026 04:30:23 +0000

is that really what people want? The fact that people say why not have 50 story concrete blocks everywhere to get more people feels like exact thing that would destroy what makes living in the city nice... Tenement housing sucked, why add thousands of people to crammed parts of city. We should be incentivizing sprawl and better transportation.

New comment by bfeynman in "Cutting inference cold starts by 40x with LP, FUSE, C/R, and CUDA-checkpoint"

bfeynman — Mon, 18 May 2026 19:28:54 +0000

probably just AI slop and using wrong semantics, they mean speedup ratio.

New comment by bfeynman in "Bun Rust rewrite: "codebase fails basic miri checks, allows for UB in safe rust""

bfeynman — Fri, 15 May 2026 18:48:11 +0000

it's more straightforward to write safe rust when rust owns everything, In real world you often are interfacing with underlying libs or systems etc, which you need to treat as invariants but also handle yousrelf manually to make guarantees to compiler. unsafe exists in tons of codebases it's just you have to make sure you encapsulate it properly, which is what this bug is.

New comment by bfeynman in "Quack: The DuckDB Client-Server Protocol"

bfeynman — Fri, 15 May 2026 18:02:39 +0000

Very hyped for this and updates. Have been using my own workarounds for a while with own WAL things and then sort of generating snapshots which with duckdb is so cheap was simpler than really implementing concurrent writes and mutations but this will make it so much easier.

New comment by bfeynman in "Launch HN: Voker (YC S24) – Analytics for AI Agents"

bfeynman — Tue, 12 May 2026 18:29:41 +0000

do you have experience as PMs? Looking at website, it looks like you just use llms to guess what categories are? Seems like trap for garbage in garbage out. Otherwise you would need someone technical to figure out how to setup the proper KPI monitoring things...

New comment by bfeynman in "The 'Hidden' Costs of Great Abstractions"

bfeynman — Mon, 04 May 2026 15:47:09 +0000

A lot of software practices are not based on engineering but rather psychology, and based on the fact that software costs $$$ to develop and maintain. There has been a complete paradigm shift with AI and it has underlined all of this. It's plausible there is no need (in high level programming at least) for any sorts of best practices, design patterns, code hygiene because not only will persons not even be looking at it, but AI can just rewrite an entire service if it needs to add a small feature and refactor everything, nothing needs to be written to be scalable or extensible if this cost becomes free. This is getting closer and closer to reality every month.

New comment by bfeynman in "We gave an AI a 3 year retail lease and asked it to make a profit"

bfeynman — Thu, 16 Apr 2026 21:28:50 +0000

How is it not analogous to data leakage? The claim is that the system works autonomously, or at minimum could, but there is effectively signal via human in the loop feedback. That's leakage into test time evaluation. Also the coding analogy is malappropriated, in that the llm is using its own signals autonomously in the environment. Using a kalman filter on a ICBM with its own sensors is analogous to the coding agent and is autonomous. A system where a human is course correcting based on signals/sensor data is what's presented here, that is not autonomous.

New comment by bfeynman in "We gave an AI a 3 year retail lease and asked it to make a profit"

bfeynman — Thu, 16 Apr 2026 19:34:41 +0000

that is ... not correct? This is classic example of data leakage, the yes/no things are signals feeding back to the model influencing (and here, basically guiding) future decisions.

New comment by bfeynman in "We gave an AI a 3 year retail lease and asked it to make a profit"

bfeynman — Thu, 16 Apr 2026 17:44:36 +0000

I feel bad that people have to read this. It's complete puffery, made up for clicks, and the biggest thing is the pure bravado with which a company says, "Hey, let's just waste a ton of money, all for a potential blog and marketing piece." This is not really automated in any fashion. I was dubious at first, but then I saw the screencaps showing the devs interacting with Luna via a Slack workflow with a human in the loop — meaning they're literally just proxying their own behavior through an LLM. This is no different than anyone who consults AI for any decision with context. To get even more technical on the fallacy: this is not automation, as there is data leakage at every step where there is a human in the loop. A broken clock is right twice a day; an LLM could cycle through 100 guesses to pick a number, but don't market that as an oracle. Aside from that, you could just look at the pictures and context (retail in SF) and assume making a profit here would be near impossible. An actual AI ceo would probably have immediately cancel the lease.

New comment by bfeynman in "Next Grok model training with 10T parameter model"

bfeynman — Wed, 08 Apr 2026 18:11:41 +0000

Isn't what the leading labs are currently chasing after is not pretraining and massive parameters but enriched and deep fine tuning and post training for agentic tasks/coding? MoE with just new post training paradigms lets smaller models perform quite well, and much more pragmatic to scale inference with. Given that, this choice seems super odd, as the frontier labs seem to stay neck and neck, and I don't even see Grok being used in any benchmarks because of how poorly it performs