Hacker News: olmo23

New comment by olmo23 in "Anthropic's Safety Superpower"

olmo23 — Mon, 15 Jun 2026 11:27:18 +0000

> no nationality controls in place

Not for now, but how long before we have KYC regulations concerning LLMs?

New comment by olmo23 in "MTG Bench: Testing how well LLMs can play Magic"

olmo23 — Fri, 12 Jun 2026 08:46:52 +0000

There is also this Matt Parker video about MTG, in which he explores a specific three-card combination that produces an ungodly amount of creature tokens.

https://www.youtube.com/watch?v=x3dE-NJ1UDQ

New comment by olmo23 in "OpenAI mulls slashing prices as it competes with Anthropic for users"

olmo23 — Thu, 11 Jun 2026 06:14:21 +0000

ZDR = Zero Data Retention

New comment by olmo23 in "Albania Is Not for Sale: Kushner's $4B Resort Triggers'Flamingo Revolution'"

olmo23 — Tue, 09 Jun 2026 15:36:06 +0000

What about drones though

New comment by olmo23 in "Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes"

olmo23 — Thu, 04 Jun 2026 11:34:08 +0000

Perhaps, but I'm not sure my boss would appreciate that though.

New comment by olmo23 in "When AI Crosses the Line: The Matplotlib Incident"

olmo23 — Mon, 01 Jun 2026 13:37:14 +0000

> And perhaps the people who built and deployed the autocomplete and the connection as well.

I disagree. IMO it's the person who connects the LLM to the button who bears the responsibility of the workings of the resulting contraption.

New comment by olmo23 in "The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin"

olmo23 — Fri, 29 May 2026 09:45:31 +0000

Since my employer pays for it, I just select the latest and greatest.

New comment by olmo23 in "Incident with Pull Requests, Issues, Git Operations and API Requests"

olmo23 — Wed, 27 May 2026 12:58:22 +0000

What really grinds my gears is how easy it is to get better designs out of LLMs. But if you don't ask, you get the default.

New comment by olmo23 in "Nobody cracks open a programming book anymore"

olmo23 — Tue, 26 May 2026 12:11:14 +0000

> Funnily enough, people on HN often do not consider this an issue, like at all...

I didn't have a problem with it when it was Aaron Swartz, not sure why I should have a problem with it when others do it.

New comment by olmo23 in "What Do Gödel's Incompleteness Theorems Mean?"

olmo23 — Thu, 21 May 2026 14:09:59 +0000

Also check out https://en.wikipedia.org/wiki/Rice%27s_theorem

basically generalized the halting problem to arbitrary semantic properties.

New comment by olmo23 in "Anthropic is expanding to Colossus2. Will use GB200"

olmo23 — Thu, 21 May 2026 09:55:35 +0000

I basically open up a new conversation, copy paste a link to Theo's latest video and ask it to summarize his yapping :p

New comment by olmo23 in "Anthropic is expanding to Colossus2. Will use GB200"

olmo23 — Thu, 21 May 2026 06:41:53 +0000

My go-to usecase for Gemini is summarizing Youtube tech-influencers.

New comment by olmo23 in "GTFOBins"

olmo23 — Tue, 28 Apr 2026 07:31:34 +0000

In certain circumstances, they might be :-)

But you can't "hack a server" using just these techniques: they would be a (small) part of a chain of exploits.

New comment by olmo23 in "Talkie: a 13B vintage language model from 1930"

olmo23 — Tue, 28 Apr 2026 07:29:04 +0000

This is an active area of research. Demis Hassabis proposed training a model with a strict knowledge cutoff before 1915, and seeing whether it can independently arrive at general relativity.

New comment by olmo23 in "Kernel code removals driven by LLM-created security reports"

olmo23 — Wed, 22 Apr 2026 13:43:59 +0000

We are there. This is pretty much the reason why Mythos isn't being released publically.

New comment by olmo23 in "AI could be the end of the digital wave, not the next big thing"

olmo23 — Mon, 13 Apr 2026 14:22:50 +0000

I don't think it would take very long to regain those skills either.

New comment by olmo23 in "They See Your Photos"

olmo23 — Mon, 13 Apr 2026 14:09:08 +0000

Cropping doesn't necessarily remove exif. It doesn't even always remove the original pixels (simply setting cropTop and cropLeft and similar fields).

My mind was blown when I saw rainbolt uncrop a picture.

Anyways as mentioned elsewhere: when I tried it the vision api was overloaded but I still received the location data. And it was from a picture taken inside my car (no landmarks or horizons visible).

New comment by olmo23 in "They See Your Photos"

olmo23 — Mon, 13 Apr 2026 13:54:24 +0000

I think it uses exif data for that, because when I tried it did show my location but the vision api was overloaded, so nothing more showed up.

New comment by olmo23 in "A compelling title that is cryptic enough to get you to take action on it"

olmo23 — Sat, 11 Apr 2026 13:42:49 +0000

A short reply without much substance

New comment by olmo23 in "Gone (Almost) Phishin'"

olmo23 — Thu, 02 Apr 2026 09:47:58 +0000

I told my parents: if they are ever called by anyone, to tell them "now is not a good time, please give me a case number and I'll call back when I do have the time."

And then, this is important, look up the number for the customer service hotline online.

I feel like this is a simple solution that works 100% of the time.