Hacker News: polotics

New comment by polotics in "Eunice is almost here"

polotics — Sun, 25 Oct 2026 20:54:09 +0000

though just as timely

New comment by polotics in "[dead]"

polotics — Sat, 24 Oct 2026 20:46:35 +0000

followed by the hilarious "and a small team with a few hours to finish the job before anyone wakes up"... it's a silent helicopter, everyone is still sleeping!

New comment by polotics in "I don't understand quantum physics (2018)"

polotics — Sat, 24 Oct 2026 10:02:59 +0000

me neither! How does the Einstein Rosen Podosky Paradox work? can the unknown that's being collapsed not be a pre-determined value that therefore is not traveling faster than the speed of light?

New comment by polotics in "Elon Musk Acquires $1B Stake in CNN's Parent Company"

polotics — Sat, 24 Oct 2026 08:16:45 +0000

you're welcome to develop the idea further as you see fit!

New comment by polotics in "Elon Musk Acquires $1B Stake in CNN's Parent Company"

polotics — Sat, 24 Oct 2026 05:35:16 +0000

Not quite: in the timeline it's the next one that's for life, a neuralink-plugged CNN-says-he's-alive heir that will then be `alive` until humanity's extinction.

just kidding

New comment by polotics in "I Became a Citizen of Japan"

polotics — Sun, 18 Oct 2026 06:13:49 +0000

can you clarify the meaning of the word "real" here? are you ok in your reality? don't be afraid to ask for help!

New comment by polotics in "Overdose Offsets: Why Every Bag of Fentanyl I Sell Is Net Life-Positive"

polotics — Thu, 15 Oct 2026 08:08:56 +0000

I disagree there is still a whiff of non-human-ness in the output, some mechanistic "je ne sais quoi" in the posts. Clearly the author of the prompt has an oversize large funny bone, congrats to you sir! (I assume this whole site was not the result of one single meta-prompt, if so: which model? Please!)

New comment by polotics in "How to engineer an AI harness for game development"

polotics — Tue, 13 Oct 2026 18:32:56 +0000

keyword here is "small" maybe?

New comment by polotics in "We have LLMs now. Why are the docs still wrong?"

polotics — Fri, 09 Oct 2026 13:50:01 +0000

the article's prose being riddled with LLM-isms kind of answers the question in a poetic-justice kind of way...

New comment by polotics in "Bender: AI's 'Existential' Risk Is 'Fake' [video]"

polotics — Wed, 07 Oct 2026 16:17:19 +0000

Yes! I was so looking forward to that! Please please someone with tokens to burn convert that to Futuramaaaah!!!!

New comment by polotics in "Don't be fooled–LLMs don't reason"

polotics — Fri, 02 Oct 2026 18:18:26 +0000

Friendly reminder that y'all could be reading this, or you could take things a bit more seriously and check out prof. Robert Sapolsky's excellent (2024) series of lectures on human behavioural biology. (eg. on YouTube, unless your're a lucky californian) Anyone who wants to venture an opinion about this or that reasoning should be interested in finding out about their own hardware & software first, right?

New comment by polotics in "Benchmarking retrieval for agents on messy real-world company knowledge"

polotics — Fri, 02 Oct 2026 17:59:57 +0000

ok I bite!

What is your opinion of:

- BEAM

- MemEval

- MemoryArena

- any other you can find on eg. HuggingFace

How much value do you see in Karpathy's gist on the Episodic/Procedural/Semantic split (aka. btw. Doxa/Koine/Gnosis) If you do see value what parts of your design matches?

As you are using real company data, will you at least make one step towards reproducibility by publishing some extraction pipeline so other companies can run their own comparison?

New comment by polotics in "AI godfather Yann LeCun: Anthropic CEO deluded, doesn't understand cybersecurity"

polotics — Fri, 02 Oct 2026 09:17:59 +0000

Mmh he is also known for clear thinking and teaching. As for past/present/future worth, I think you're just being ageist, can you show your work maybe, what are you working on to get to that "past accomplishments" burn some day?

New comment by polotics in "DeepSeek Harness Desktop for macOS and Windows"

polotics — Fri, 02 Oct 2026 06:33:35 +0000

how could the authors of frontierharness.org not see they used the exact same symbol for Claude Code and OhMyPi?

New comment by polotics in "Gemini 4 Argon"

polotics — Wed, 30 Sep 2026 20:17:37 +0000

traces or it didn't happen!

New comment by polotics in "Gemini 4 Argon"

polotics — Wed, 30 Sep 2026 20:16:01 +0000

Mmh ok. How much theoretical speed or 'intelligence' gain is realized by allowing reasoning to occur in some inscrutable intermediate representation? Has this been actually tested, how much is it slowing them down, and compared to whom exactly?

New comment by polotics in "September 2026: The world today, as seen by one Polish guy"

polotics — Wed, 30 Sep 2026 15:53:45 +0000

Have you noticed the small detail about the application of the sanctions bill that has been signed? Please tell us when you see something out of this bill, and the effect it has on whom.

New comment by polotics in "UnoDOS"

polotics — Wed, 30 Sep 2026 06:09:32 +0000

Sorry but this language is so painful to read:

"What has actually been verified is a separate question from what builds, and PLATFORMS.md is the ledger that answers it:"

...that there is no way I'm going to try unodos.

New comment by polotics in "Netanyahu claims ability to hack any iPhone and plant false evidence"

polotics — Tue, 29 Sep 2026 16:04:14 +0000

Theory number one after catching up with some Israeli friends: he is clutching at straws trying to do preemptive damage control about the evidence that is going to come out about him...

Namely that on top of the documented fact that he turned a blind eye to the funding channels that were strengthening Hamas, because he likes his enemies to be radical non-negotiators, he also was informed about October 7th but decided it would be expedient to let the horror happen as he thought he could use it to strengthen his power.

Btw. I could not bring myself to watch Fauda season 5.

New comment by polotics in "Show HN: Raven – The harness of harnesses, built for RSI"

polotics — Tue, 29 Sep 2026 14:12:57 +0000

Well for starters the use of the term "Recursive" is very dubious.

This is as far as the eye can see all very iterative, there is no tail-call or anything fancy, it is loops. Also "Recursive" I feel kind of tries to imply the LLM's weights are being pushed around in some feedback, because to recurse you have to invoke the thing you're recursing into at its very start right? That would be reinforcement probably, and there is none of that in any such RSI so far, at least not public. Please someone contradict me with examples.

So please: "ISI" for iterative self improvement is fine. Also "ISI" does not fit the (outdated!) Vernon Vinge "singularity" trope, and that is a good thing!