HackerTrans
TopNewTrendsCommentsPastAskShowJobs

maxutility

1,143 karmajoined 7 jaar geleden

Submissions

Accidental Anonymity

macwright.com
3 points·by maxutility·16 dagen geleden·0 comments

Updated Anthropic Privacy Policy (effective 7/8/26)

anthropic.com
2 points·by maxutility·24 dagen geleden·0 comments

The Fable takedown: what happened when, what we know, what comes next

thezvi.substack.com
3 points·by maxutility·26 dagen geleden·1 comments

Ask HN: Favorite prompts for improving LLM output?

2 points·by maxutility·29 dagen geleden·2 comments

Cyber lack of security and AI governance

thezvi.substack.com
1 points·by maxutility·2 maanden geleden·0 comments

Nextpad++ feels like a fever dream

daringfireball.net
3 points·by maxutility·2 maanden geleden·2 comments

Google tools for customizing searches

cardcatalogforlife.substack.com
79 points·by maxutility·2 maanden geleden·20 comments

The Secret Weapon Against AI Dominance

theatlantic.com
2 points·by maxutility·2 maanden geleden·0 comments

Near-instantly aborting the worst pain imaginable with psychedelics

psychotechnology.substack.com
5 points·by maxutility·3 maanden geleden·0 comments

I can never talk to an AI anonymously again

theargumentmag.com
5 points·by maxutility·3 maanden geleden·0 comments

Google has a secret reference desk

cardcatalogforlife.substack.com
23 points·by maxutility·3 maanden geleden·6 comments

The decade long feud shaping the future of AI

wsj.com
1 points·by maxutility·3 maanden geleden·0 comments

Refine: AI-Powered Peer Review

refine.ink
6 points·by maxutility·3 maanden geleden·1 comments

Politics and the English Language, January 2026 edition

daringfireball.net
3 points·by maxutility·5 maanden geleden·0 comments

The most popular blogs of Hacker News in 2025

refactoringenglish.com
3 points·by maxutility·6 maanden geleden·0 comments

The 26 Most Important Ideas for 2026

derekthompson.org
4 points·by maxutility·6 maanden geleden·0 comments

The Problem with Letting AI Do the Grunt Work

theatlantic.com
3 points·by maxutility·6 maanden geleden·0 comments

AI is mastering language. Should we trust what it says? (2022)

nytimes.com
3 points·by maxutility·7 maanden geleden·1 comments

Heiliger Dankgesang: Reflections on Claude Opus 4.5

hyperdimensional.co
3 points·by maxutility·7 maanden geleden·0 comments

Andrej Karpathy on improved Tesla FSD performance under HW4

twitter.com
1 points·by maxutility·8 maanden geleden·1 comments

comments

maxutility
·4 dagen geleden·discuss
When the entire article reads like it’s written by an LLM —- ie the author couldn’t be bothered to write their thoughts out themselves —- it’s hard to give it the benefit of the doubt that it’s worth reading closely.
maxutility
·18 dagen geleden·discuss
Dupe: https://news.ycombinator.com/item?id=48633594
maxutility
·24 dagen geleden·discuss
[dead]
maxutility
·26 dagen geleden·discuss
Zvi’s writing style is a bit of an acquired taste, and he has strong opinions which can fall outside of the mainstream, but there’s no one in the industry more well read and known for doing the reading and documenting the nitty gritty of weekly developments in the world of AI models, discourse, and policy.

Even many who disagree strongly with his worldview find his notes and references to be an irreplaceable resource for understanding the fast moving world of AI.
maxutility
·2 maanden geleden·discuss
lots of useful Google search tricks and syntax all in one place. I already knew many of these. But verbatim mode is new to me and addresses a major complaint I’ve had about increasingly fuzzy semantic search.
maxutility
·3 maanden geleden·discuss
Here’s John H Cochrane (the grumpy economist, Substack) praising refine. This is the post that first clued me into the service: https://www.grumpy-economist.com/p/refine

Here’s the Wikipedia for Cochrane: https://en.wikipedia.org/wiki/John_H._Cochrane

I haven’t tried refine myself, but from second hand reports it seems potentially well designed and useful.
maxutility
·4 maanden geleden·discuss
It seems like Disney’s departure from its business “deal” with OpenAI is newsworthy distinct from the Sora closure. Six months ago Sam Altman was announcing massive deals left and right, now nearly all of the impressive deals have fallen through or been scaled dramatically back.
maxutility
·4 maanden geleden·discuss
I found Sam's early 2015 posts on machine superintelligence and regulation [1] [2] to be even more interesting in hindsight, given OpenAI's accelerationist bent of late, OpenAI president Greg Brockman's lobbying efforts against AI regulation, and frequent accusations of attempted regulatory capture.

[1] https://blog.samaltman.com/machine-intelligence-part-1 [2] https://blog.samaltman.com/machine-intelligence-part-2

Sam's recommendations at the time include: 1) Provide a framework to observe progress… 2) Given how disastrous a bug could be, require development safeguards to reduce the risk of the accident case. For example, beyond a certain checkpoint, we could require development happen only on airgapped computers…, require that certain parts of the software be subject to third-party code reviews, etc. 3) Require that the first SMI developed have as part of its operating rules that a) it can’t cause any direct or indirect harm to humanity (i.e. Asimov’s zeroeth law), b) it should detect other SMI being developed but take no action beyond detection, c) other than required for part b, have no effect on the world. … 4) Provide lots of funding for R+D for groups that comply with all of this, especially for groups doing safety research. 5) Provide a longer-term framework for how we figure out a safe and happy future for coexisting with SMI…

Also, in his acknowledgments he gives the greatest thanks to onetime partner, now rival, Dario Amodei.
maxutility
·9 maanden geleden·discuss
Arizona is permanent standard time rather than permanent DST, and is thus unaffected by the permanent-DST winter mornings issue.
maxutility
·9 maanden geleden·discuss
Previous discussion of the offending claim: https://news.ycombinator.com/item?id=45633482

An OpenAI researcher tweeted: “ Using thousands of GPT5 queries, we found solutions to 10 Erdős problems that were listed as open: 223, 339, 494, 515, 621, 822, 883 (part 2/2), 903, 1043, 1079.

Additionally for 11 other problems, GPT5 found significant partial progress that we added to the official website: 32, 167, 188, 750, 788, 811, 827, 829, 1017, 1011, 1041. For 827, Erdős's original paper actually contained an error, and the work of Martínez and Roldán-Pensado explains this and fixes the argument.”

This was taken (out of context?) to be claiming ChatGPT solved the open problems, when in fact it “just” found them through a literature review. (Though an earlier tweet in the same thread made the literature review interpretation more explicitly)

The ensuing controversy around whether there was false hype buried a potentially significant demonstration of LLMs’ ability to unlock lost and forgotten knowledge in a way Sebastien explains and makes the case here as being a big deal.