Hacker News: loglog

New comment by loglog in "A misalignment of AI in mathematics"

loglog — Fri, 11 Sep 2026 20:27:22 +0000

> Mathematics is about discovering and understanding the logical implications of assumed axioms under various inference rules.

That's like saying that programming is about producing valid programs in various programming languages.

New comment by loglog in "Go grandmaster Shin defeats AI KataGo with a two-stone handicap"

loglog — Thu, 03 Sep 2026 20:59:09 +0000

Distillation of our blessed models is no fair!

New comment by loglog in "GUIs should be fully keyboard-driven"

loglog — Sat, 29 Aug 2026 10:07:24 +0000

> Also in a TUI you never ever have that issue that some text isn't copy/pastable like some dialogue boxes

May I present OpenCode, which somehow managed to break copy from terminal: https://github.com/anomalyco/opencode/issues/4283

New comment by loglog in "GLM-5.3 is now open-weight"

loglog — Fri, 28 Aug 2026 20:53:44 +0000

Self-hosting won't protect you from getting locked out of a closed weights model, because you can't self-host it even if you have the hardware.

New comment by loglog in "GLM-5.3 is now open-weight"

loglog — Fri, 28 Aug 2026 20:50:59 +0000

If they "should" cost 4k in the sense of marginal cost, then you will be spending more running the same at home, because your home hardware will always be less efficient.

New comment by loglog in "Show HN: The load-bearing vocabulary of Claude"

loglog — Fri, 28 Aug 2026 08:52:23 +0000

As of now, there are only 2 natural language clusters in the partition: Claudish and Spanish/French. From a first glance, the other clusters seem centered on technologies (lots of command line flags and abbreviations among the "most representative" words). If the goal is to investigate English language usage trends, it might be better to cluster based only on English words.

New comment by loglog in "The turbulent AI era is here"

loglog — Thu, 27 Aug 2026 21:15:41 +0000

Their suggestion to tax tokens extends the problem of taxing wages to a new type of production. The real solution is, of course, taxing the capital, but you will never hear that from them.

New comment by loglog in "If you are asking for human attention, demonstrate human effort"

loglog — Sat, 13 Jun 2026 08:53:15 +0000

The bots can also write Rust instead of C++, doing away with the arcane nonsense accumulated by that legacy language. SCNR.

New comment by loglog in "If you are asking for human attention, demonstrate human effort"

loglog — Sat, 13 Jun 2026 08:47:23 +0000

Bureaucracy gives you leverage against slop. Review seriously, but limit the time that you spend. This will stall the slop. When the culprit complains, tell your boss "I spend X hours per week on reviews. If you need more throughput, the PRs quality needs to improve."

New comment by loglog in "If you are asking for human attention, demonstrate human effort"

loglog — Sat, 13 Jun 2026 08:38:58 +0000

Writing and reading design documentation can be slower than pair programming. On the other hand, info about code design also belongs into inline documentation or commit messages (in this order of preference), so the effort might not be wasted.

New comment by loglog in "Programmers will document for Claude, but not for each other"

loglog — Fri, 05 Jun 2026 13:39:52 +0000

It's not just documentation. Everything that makes programming easier is now suddenly valued, because wasting tokens is obviously worse than wasting your employees' time.

New comment by loglog in "How to convert between wealth and income tax"

loglog — Fri, 22 May 2026 22:10:59 +0000

This claim is plain malicious. Of course falling asset prices would be excellent for the median person, since they would be less extremely priced out of everywhere. This is one of the central benefits of a wealth tax.

New comment by loglog in "A more efficient implementation of Shor's algorithm"

loglog — Thu, 14 May 2026 07:38:36 +0000

> For that kind of implementation, ...

This is the key point, what is the meaning of "zero knowledge" here? It seems that you need to know something about the implementation, even if it is not the full implementation. Compare this to a zero knowledge proof that you have, say, a factorization gadget, which works by you running the gadget on adversarial input, thus convincing the adversary that you can factor any of their integers. That discloses no implementation details of your factorization gadget, which can be an efficient classical algorithm, a quantum computer, or a phone line to God.

New comment by loglog in "If AI writes your code, why use Python?"

loglog — Tue, 12 May 2026 15:31:16 +0000

The list of tools that Pythonheads present as a definite solution to their problems changes every year, yet the results are still far behind Rust/Scala/Kotlin/C#.

New comment by loglog in "Poland is now among the 20 largest economies"

loglog — Sat, 09 May 2026 09:09:21 +0000

Surpassing DB's punctuality is the first large-scale example of the "Overtaking without catching up" East German slogan coming true.

New comment by loglog in "The Disappearance of the Public Bench"

loglog — Fri, 08 May 2026 20:07:11 +0000

The "fix everything" button is abolishing zoning laws, and its aggregate cost is negative. Aggregate cost is not the issue preventing problems from being solved.

New comment by loglog in "Mojo 1.0 Beta"

loglog — Fri, 08 May 2026 19:46:24 +0000

> on top of a well designed language constructed over past language design experience

While I believe that Chris Lattner is a great compiler designer, his language design record has been less stellar. Swift bidirectional type inference for instance feels like it was implemented because they had a compiler algorithm that they wanted to use, rather than a genuine need, and is just a completely avoidable problem. Trying to make a HPC language that is also Python compatible was doomed from the start. Hopefully the damage from going into this direction will remain limited.

New comment by loglog in "Mojo 1.0 Beta"

loglog — Fri, 08 May 2026 19:36:45 +0000

Mojo is NOT Python compatible (although they initially wanted it to be). So they got all downsides without the upsides.

New comment by loglog in "A more efficient implementation of Shor's algorithm"

loglog — Sun, 03 May 2026 13:26:53 +0000

How is it possible to provide a zero knowledge proof that their circuit works for large problem instances if there is no efficient way to run or simulate the circuit with the required instance size?

New comment by loglog in "“Why not just use Lean?”"

loglog — Mon, 27 Apr 2026 17:56:26 +0000

"I believe that almost anything that has been formalised today in any system could have been formalised in AUTOMATH. Its main drawbacks were its notation, which really was horrible, and its complete lack of automation. Proofs were long and unreadable." That's like saying that anything that could be programmed today in your modern language of choice could have been programmed 50 years ago in assembly. Technically yes, economically no.