Hacker News: shwaj

New comment by shwaj in "Research acceleration: The view inside OpenAI"

shwaj — Mon, 07 Sep 2026 02:12:22 +0000

I think you’re conflating the direction of definition with the direction of evaluation.

Compare the similarity of:

  AI(n) = improve(AI(n-1))

With:

  Fib(n) = Fib(n-1) + Fib(n-2)

The latter is a classic example of recursion. So why isn’t the former?

Edit: formatting

New comment by shwaj in "Research acceleration: The view inside OpenAI"

shwaj — Sun, 06 Sep 2026 19:41:10 +0000

“Recursive” is a reasonable term because the generation N AIs will train the Generation N+1 AIs. The term “iterative” doesn’t reflect this nuance as well IMO.

New comment by shwaj in "QBittorrent breaks out of sandbox to commit crimes"

shwaj — Sun, 06 Sep 2026 18:27:07 +0000

The risks of open models are real.

However, I’d feel better about ceding control of AI to the government if they didn’t seem so intent on building an apparatus for surveillance/control.

Your “furry fanfic” is a somewhat pathetic scarecrow; this is a complicated topic.

New comment by shwaj in "QBittorrent breaks out of sandbox to commit crimes"

shwaj — Sun, 06 Sep 2026 18:18:48 +0000

Neuromancer came out in 1984, over 40 years ago, so that covers the broad concept. Or Neo in The Matrix: “Programs hacking programs… why?”

For non-fiction you could look at Nick Bostrom’s Superintelligence: Paths, Dangers, Strategies (2014).

More recently, see Cade Metz’s NYT profile of Geoffrey Hinton, “The Godfather of A.I. Leaves Google and Warns of Danger Ahead” (May 1, 2023).

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

shwaj — Fri, 04 Sep 2026 07:23:09 +0000

I was under the impression that Go also uses Elo, then I did a bit of cursory research and discovers that it varies.

Two major federations are American Go Association (AGA) and European Go Federation (EGF). EGF uses an Elo-inspired update rule since 2021. AGA uses a quite-different Bayesian system without pairwise update; they provide a paper and a C++ reference impl.

Asian countries don't bother with such numeric ratings. Instead, rankings are titles which are won through tournament promotion structures (sounds similar to Sumo to me).

Interesting, because I always thought that it was more "apples to apples", and that the higher upper limits of Go rankings was somehow indicative of the higher "dynamic range" of the game compared to chess. For example, if Elo were applied to basketball, what would the Elo of Lebron James be compared to a playground hooper (leaving aside that 1-on-1 isn't the best part of Lebron's game)... would it be higher or lower than Magnus Carlsen in chess? I don't have an intuition.

New comment by shwaj in "Apple caught off guard by AI demand for Mac Mini and Mac Studio"

shwaj — Tue, 01 Sep 2026 15:38:03 +0000

It was a joke referencing the typo in the message they responded to.

New comment by shwaj in "Carmack: Has early Scratch experience led to fulfilling careers?"

shwaj — Tue, 01 Sep 2026 02:18:29 +0000

I was housemates with a couple with a kid. The kid learned Scratch first, and made some surprisingly sophisticated apps in that semester (or two).

By now he’s moved on to textual languages (not yet old enough for a career), but Scratch was engaging and gave him a good start.

New comment by shwaj in "How much of HN is AI?"

shwaj — Tue, 25 Aug 2026 19:07:47 +0000

I don't know about the article about Apple News scams, but how is an article about Waymo publishing a model for self driving cars not about AI?

New comment by shwaj in "Moon (2024)"

shwaj — Tue, 25 Aug 2026 01:07:55 +0000

It's not plagarism, Ciechanowski didn't invent this presentation form. His presentations are examples of "active essays", a term coined by Alan Kay no later than 2006.

(I don't know whether Alan invented the concept itself, I'm not aware of him claiming that)

See: https://ieeexplore.ieee.org/abstract/document/5350160

New comment by shwaj in "CIA funding helped keep NeXT afloat in the 80s"

shwaj — Thu, 20 Aug 2026 16:43:07 +0000

Your tone implies that the same people who freak out about reasonable gov intervention, are also (hypocritically) okay with the shenanigans being discussed.

Who are these straw men you refer to?

New comment by shwaj in "Amazonian civilization had estimated 3M people in 3% of forest area"

shwaj — Tue, 04 Aug 2026 04:02:01 +0000

You could be more charitable. They did say “or” which you omitted in your quote, and also “who knows”. There are possibilities too. In my view, the primary contribution of that comment was to clarify the time frame.

New comment by shwaj in "AFC Stands in Solidarity with UEFA and Concacaf to Protect the FIFA World Cup"

shwaj — Fri, 31 Jul 2026 19:51:58 +0000

I have to assume that was sarcasm.

New comment by shwaj in "Claude Opus 5"

shwaj — Fri, 24 Jul 2026 18:16:32 +0000

So highlight both then.

New comment by shwaj in "Claude Opus 5"

shwaj — Fri, 24 Jul 2026 17:38:53 +0000

It would still be the best model per dollar if the score was 2% lower instead of 0.1% lower. Would it be ok to still give it the highlight color then?

How big of a lie is too big? Especially when no lie needed to be told at all: many including myself would have noticed the tiny 0.1% deficit and been suitably impressed by the Opus 5 result.

I’ll admit this is a small deception by today’s standards. I’m one of those who believes in truth for truth’s sake.

Edit: typo

New comment by shwaj in "The arguments against open source AI are bad"

shwaj — Fri, 24 Jul 2026 17:31:06 +0000

I didn’t grow up with that hobby. The point stands anyway.

Cracking a binary circumvents authorization to provide full access to functionality the program already has. That’s very different from modifying the program to add new features.

New comment by shwaj in "Claude Opus 5"

shwaj — Fri, 24 Jul 2026 17:19:44 +0000

I like how they highlighted Opus 5 as the best for “Agentic Coding” even though the number is slightly lower than Fable. Close enough for marketing, I guess!

New comment by shwaj in "The arguments against open source AI are bad"

shwaj — Fri, 24 Jul 2026 00:20:13 +0000

Equating an open weight model to a closed source binary is ludicrous. An open weight model can be fine tuned, for example. How easy is it to modify a closed source binary for your purposes?

New comment by shwaj in "OpenAI and Hugging Face address security incident during model evaluation"

shwaj — Wed, 22 Jul 2026 02:30:04 +0000

They avoid the full basilisk treatment.

New comment by shwaj in "Kimi K3: Open Frontier Intelligence"

shwaj — Fri, 17 Jul 2026 13:59:37 +0000

Since we’re hair splitting: that’s not what bike shedding means. :-D

New comment by shwaj in "Modern decor may be straining people's brains"

shwaj — Sat, 11 Jul 2026 20:14:13 +0000

“Striped office floors. Flickering lights. Walls covered in repetitive geometric patterns.”

Not sentences. AI slop.