Hacker News: ttpphd

New comment by ttpphd in "Mayo Clinic's secret weapon against AI hallucinations: Reverse RAG in action"

ttpphd — Sat, 15 Mar 2025 16:58:30 +0000

If LLMs were good at summarization, this wouldn't be necessary. Turns out a stochastic model of language is not a summary in the way humans think of summaries. Thus all this extra faff.

New comment by ttpphd in "Stanford scientists discover new Ozempic alternative with RegEx"

ttpphd — Sat, 08 Mar 2025 14:26:49 +0000

You didn't believe it then, but you better believe it now: RegEx is the final boss of both code and science.

New comment by ttpphd in "An update on Mozilla's terms of use for Firefox"

ttpphd — Sat, 01 Mar 2025 13:16:22 +0000

Exactly.

Why is my browser serving me advertising in the first place? Because Mozilla is an advertising company now.

New comment by ttpphd in "DeepSeek Open Infra: Open-Sourcing 5 AI Repos in 5 Days"

ttpphd — Sat, 01 Mar 2025 12:32:22 +0000

This is just denialism.

New comment by ttpphd in "OpenAI expands Deep Research to all paying ChatGPT users"

ttpphd — Wed, 26 Feb 2025 02:11:49 +0000

It's amazing that no one really cares. Would you have believed that two years ago?

New comment by ttpphd in "Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs [pdf]"

ttpphd — Wed, 26 Feb 2025 01:18:09 +0000

This isn't what I meant but thanks anyway.

New comment by ttpphd in "Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs [pdf]"

ttpphd — Tue, 25 Feb 2025 22:30:10 +0000

This is where my thoughts went too. I see no reason to speculate about this in the absence of clear and persuasive comparison examples with other fine tuning content.

New comment by ttpphd in "DOGE has 'god mode' access to government data"

ttpphd — Thu, 20 Feb 2025 18:40:21 +0000

This is wrong and naive.

https://talkingpointsmemo.com/edblog/doge-dives-into-core-na...

"DOGE currently has far deeper and far more extensive access to U.S. government computer systems — and is far deeper into the national security space — than is conceivably necessary for anything related to their notional brief and goals."

New comment by ttpphd in "Accelerating scientific breakthroughs with an AI co-scientist"

ttpphd — Thu, 20 Feb 2025 14:29:30 +0000

Are you a scientist?

New comment by ttpphd in "DOGE has 'god mode' access to government data"

ttpphd — Thu, 20 Feb 2025 14:18:40 +0000

New comment by ttpphd in "Accelerating scientific breakthroughs with an AI co-scientist"

ttpphd — Wed, 19 Feb 2025 17:38:41 +0000

It's almost like scientists are doing something more than a random search over language.

New comment by ttpphd in "Accelerating scientific breakthroughs with an AI co-scientist"

ttpphd — Wed, 19 Feb 2025 17:37:43 +0000

Witness the pure arrogance of tech bros

New comment by ttpphd in "Accelerating scientific breakthroughs with an AI co-scientist"

ttpphd — Wed, 19 Feb 2025 16:52:15 +0000

Are you a scientist?

New comment by ttpphd in "LLMs can teach themselves to better predict the future"

ttpphd — Wed, 12 Feb 2025 14:34:37 +0000

You didn't make a case for any of that. No one did. This whole discussion is just a bunch of people who have their feelings hurt when other people tell them a LLM is modeling language, not reasoning. It's so narcissistic. "My opinion on AI is criticized so I'm oppressed."

That is not how oppression and power work. That's not how discussion works. That not how Foucault's analysis of power works.

New comment by ttpphd in "LLMs can teach themselves to better predict the future"

ttpphd — Tue, 11 Feb 2025 19:38:11 +0000

It's absolutely not taboo to believe that. It's a very common belief.

Lot of game playing going on here to center on a victim narrative.

New comment by ttpphd in "LLMs can teach themselves to better predict the future"

ttpphd — Tue, 11 Feb 2025 19:29:51 +0000

"societal taboos and power structures can be revealed by noting what speech provokes the strongest reactions"

Ok I'll bite. Who is the marginalized Other?

New comment by ttpphd in "LLMs can teach themselves to better predict the future"

ttpphd — Tue, 11 Feb 2025 19:17:41 +0000

Throwing your hands up in the air like this doesn't help build a constructive case for using the word reasoning. It builds a case that words mean whatever

New comment by ttpphd in "Teen on Musk's DOGE team graduated from 'The Com'"

ttpphd — Mon, 10 Feb 2025 02:04:08 +0000

"if the president does it, it's not illegal" ok Nixon.

New comment by ttpphd in "Teen on Musk's DOGE team graduated from 'The Com'"

ttpphd — Sun, 09 Feb 2025 20:53:45 +0000

Gullibility is a thing, which is why we demand people go through channels of accountability, including security clearances.

New comment by ttpphd in "Teen on Musk's DOGE team graduated from 'The Com'"

ttpphd — Sun, 09 Feb 2025 20:51:27 +0000

Yes my main takeaway from the musk flagging discussions on here is that tech people are no where near capable of taking responsibility for their central role in this mess. If I call all my sins "politics", say discussing politics is off the table, then magically you can't say anything bad about me!