Hacker News: jbritton

New comment by jbritton in "GPT-6 Astra"

jbritton — Fri, 04 Sep 2026 01:27:35 +0000

I decided to reply to my own comment. In the video above, the initial moves are textbook. Then a position that has never been played is reached. At this point it appears to pattern match against a similar but different board and pattern matches some follow on board. The result is illegal moves and no ability to see checks, captures, threats, tactics.

Which is strange because I’m sure it could give general advice about how to play better, it just doesn’t follow the rules it can enumerate. It also doesn’t seem to have spatial awareness.

I used to think LLMs couldn’t do Fibonacci for the same reason. They could write the code but not follow it. They can now follow a procedure to generate fib numbers but it seems to be memory limited.

So I don’t know why it can track fib algo, but no chess concepts.

New comment by jbritton in "GPT-6 Astra"

jbritton — Thu, 03 Sep 2026 21:43:23 +0000

Watch a chess bot championship here: https://youtu.be/7g-jN3DTkWQ?is=HV3cdcICIRMbswQ3

Then realize LLMs have zero of what anyone would consider intelligence.

New comment by jbritton in "Scientists release biggest 2D map of the universe"

jbritton — Sat, 22 Aug 2026 00:30:43 +0000

I will add that it is said the universe is flat, but that does not mean 2d flat. Flat is a mathematical construct, which means parallel lines stay parallel and triangle angles sum to 180 degrees. So it is 3d flat at least as far as we can observe.

New comment by jbritton in "MIT neuroscientists discover the brain can reason without words"

jbritton — Fri, 14 Aug 2026 01:01:25 +0000

I’m unconvinced. I have a voice in my head that is always there flowing words while thinking. Actually being able to speak maybe different than running words through your head. I also think that maybe not every task requires the head voice, and visual pattern matching was the first thing I considered that might run an orthogonal path. But it would have to be simple. If you start doing IQ puzzles I have my doubts. I do wonder though how someone who was deaf at birth functions.

New comment by jbritton in "Devtools must be open source"

jbritton — Mon, 03 Aug 2026 20:22:11 +0000

Maybe a donate link attached too. The dollars tied to a PR could help set priority of fixes.

New comment by jbritton in "I'm 38 and I Can't Support Myself Anymore"

jbritton — Tue, 28 Jul 2026 23:16:21 +0000

Your tax dollars just go into a cesspool of fraud, corruption, waste, incompetence, distribution to cronies. Money taken and spent badly feels like theft.

New comment by jbritton in "Opus 5 is currently #1 on Artificial Analysis Intelligence Leaderboard"

jbritton — Sat, 25 Jul 2026 00:37:49 +0000

I was flagged for basically answering yes to what Claude suggested to do, which was test commands on a port for my code for tests we had been discussing. I really think it was flagged simply because the words test and port were in the prompt. Their filters are pathetically poor.

New comment by jbritton in "The unreasonable difficulty of time series forecasting"

jbritton — Tue, 21 Jul 2026 21:20:25 +0000

I recently had a conversation with Claude Fable about Renaissance Technologies. Claude says their closed fund has found a very tiny edge and their profit comes from millions of trades on that edge. Their public fund with more money doesn’t find enough edge to beat the S&P500 index.

New comment by jbritton in "The unreasonable difficulty of time series forecasting"

jbritton — Tue, 21 Jul 2026 21:10:51 +0000

This brings up the philosophical question of whether humans have free will or is everything deterministic following the laws of physics.

New comment by jbritton in "Kimi K3 just fixed 15 critical security bugs that Codex and Fable refused"

jbritton — Tue, 21 Jul 2026 03:56:55 +0000

I have a process I wrote that accepts commands on a socket. At one point Claude suggested testing it. I pretty much echoed back what Claude suggested and got blocked. All I could think was a regex matching “test” and “port” and it downgraded me. I switched back to Fable with slight rewording and was able to continue. This was for code running on my box in a Claude project folder where Claude saw me write the code. Anthropic has put zero thought into this.

New comment by jbritton in "After 7 years in production, Scarf has reluctantly moved away from Haskell"

jbritton — Sat, 11 Jul 2026 04:05:31 +0000

Well said.

New comment by jbritton in "After 7 years in production, Scarf has reluctantly moved away from Haskell"

jbritton — Sat, 11 Jul 2026 00:18:32 +0000

I write a ton of Python code. Modifying code without static types is difficult and error prone. If a function needs a new argument, all the callers have to pass it and all of their callers recursively. Maybe the calling function is just a function pointer that has been passed around and not searchable. I love Python for small tasks and tasks that are not critical if they fail.

New comment by jbritton in "TypeScript 7"

jbritton — Wed, 08 Jul 2026 23:26:46 +0000

My understanding of weak typing is about allowing automatic type conversion. “3” + 1 === “31”, 1 + “3” === 4. Strong typing doesn’t auto convert.

New comment by jbritton in "Clojure 1.13 adds support for checked keys"

jbritton — Mon, 06 Jul 2026 23:02:05 +0000

I suspect there should be some data model guideline that says if keys can be missing then values can’t be nil, or if values can be nil then keys must be present. In the famous saying there are two hard things: naming, cache invalidation, off by one errors, I think one more to add is handling missing data.

New comment by jbritton in "Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5"

jbritton — Wed, 01 Jul 2026 03:46:22 +0000

I wonder if Anthropic servers can handle the load their going to get tomorrow. Unless it’s a staged rollout

New comment by jbritton in "Claude Sonnet 5"

jbritton — Wed, 01 Jul 2026 00:24:01 +0000

I accidentally used Sonnet 5 a bit today. It seemed significantly worse to me than Opus 4.8 for software development.

New comment by jbritton in "A new bill takes aim at government pressure to silence lawful online speech"

jbritton — Fri, 19 Jun 2026 22:42:31 +0000

TOGETHER and ACTIV-6 were not good studies.

I-Tech was a good RCT and showed 75% lower mortality in severe patients, p=0.02. Only 490 people though. Clearly this deserved replicating with more people. The fact that it wasn’t is a problem.

The FDA on their website stated they did not look at the data. The statement was eventually taken down, they have never said they looked at it. The CDC looked at 6 studies and didn’t write a report.

When I mentioned method of action, I’m not talking about petri dishes.

An anecdote related to bleach, but peroxide was also mentioned. Another overlooked treatment was artemisinin, which contains an endoperoxide group that could get into cells and kill covid virus.

New comment by jbritton in "A new bill takes aim at government pressure to silence lawful online speech"

jbritton — Fri, 19 Jun 2026 20:45:35 +0000

There are multiple hypotheses for methods of action for IVM against covid that were plausible and deserved serious funding and investigation by governments. This did not happen. What did happen is the two people who brought it to the attention of congress got slandered the next day in the press and then got their careers destroyed. Anyone else who ever mentioned it also got dumped on by the press. There are over 100 studies now, many of which showed positive results. Most of the studies were small. There have been questions of the quality of many of the studies both the ones showing benefit and the ones showing no benefit. There is one well known fraudulent study and the press loves to extrapolate that to every study. The FDA has admitted to not looking at the data. The CDC and WHO did not write up any kind of analysis. They looked at a couple bigger studies that showed no result and said it’s inconclusive. No government has bothered to look into the RCTs that showed positive results. I don’t know if it works or not. It seems clear to me that profits, politics, power structures, got in the way of finding the answer. IVM is a very safe drug with the proper dosing. 300 to 500 million people in Africa take it yearly for parasites.

New comment by jbritton in "What every coder should know about Gamma Correction"

jbritton — Tue, 16 Jun 2026 01:50:49 +0000

I thought you corrected gamma in the monitor settings, not in your application code. Am I wrong.

New comment by jbritton in "Ear Training Practice Exercises"

jbritton — Thu, 11 Jun 2026 19:41:12 +0000

I never had a music class at any point in school in California.