chrisjj

27 ms·

211.
by chrisjj 2mo ago
It knows nothing of correctness or winning. It is predicting only what is most likely given its corpus.
212.
by chrisjj 2mo ago
Better title: Continue thinking of LLMs as Next-Token Predictors Because no, post training doesn't change that.
213.
by chrisjj 2mo ago
> A kill switch may be effective for an agent driven attack No doubt. But is it effective against an agent-driven attack? :)
214.
by chrisjj 2mo ago
> Anthropic has restricted access to its cyber tool Mythos on the grounds that it is too powerful to fall into the wrong hands. Too late. It's already fallen in to Anthrophic's hands.
215.
Time is running out for cyber security, warn top tech firms(bbc.co.uk)
5 points by chrisjj 2mo ago|0 comments
216.
Lords call for AI 'kill switch' powers in UK(bbc.co.uk)
7 points by chrisjj 2mo ago|3 comments
217.
by chrisjj 2mo ago
> Is Tesla still going to produce vast numbers of humanoid robots by the end of 2026? Sure, with FSD by 2030 - promise! /s
218.
by chrisjj 2mo ago
> For digital products, that would mean copyright privileges. If you sell me a (copy of a) program, I own it and need no (c) privilege to be allowed to resell it. Just like a book.
219.
by chrisjj 2mo ago
> There are some games which really do sell under "irrevocable license to play" (which, short of granting copyright privileges, is about as close as you can get to ownership for a digital product). Far from it. E.g. excludes re
220.
by chrisjj 2mo ago
> Distributed systems are hard ... when vibe-coded.
221.
by chrisjj 2mo ago
> "Gastown criticality event" is physically melting GPUs. Nice to know e-stop is functional!
222.
by chrisjj 2mo ago
And maybe that malware is one of them, playing dead.
223.
by chrisjj 2mo ago
> Elevated errors Anthropic, please hire a literate human.
224.
by chrisjj 2mo ago
> AI progress is racing along, but virtually all of the visible progress is in the realm of knowledge work, i.e. activities that can take place inside a computer. The most visible "progress" of so-called AI is in activities tha
225.
by chrisjj 2mo ago
No misunderstanding at all. "In machine learning, the term stochastic parrot is a metaphor that frames large language models as systems that statistically mimic text without real understanding." https://en.wikipedia.org
226.
by chrisjj 2mo ago
I've seen nothing to suggest it misleads anyone else.
227.
by chrisjj 2mo ago
True title: Whistleblower warns Postal Service's new mail ballot system has "potentially catastrophic problems"
228.
by chrisjj 2mo ago
An LLM is no more self-referential than a PDF reader showing its user guide.
229.
by chrisjj 2mo ago
> You run multiple debugging passes over the code until there are no more reported errors. ... leaving you to enjoy the unreported errors. Just be sure please to say "CREATED BY KNOWN UNRELIABLE SO-CALLED AI" on the start up sc
230.
by chrisjj 2mo ago
> AI is anything that can often solve inexactly specified problems I doubt you'll find many who agree. https://en.wikipedia.org/wiki/Artificial_intelligence Artificial intelligence (AI) is the capability of com
231.
by chrisjj 2mo ago
> I'll know in real time if I'm getting robbed. Yup - by checking your Claude usage meter.
232.
by chrisjj 2mo ago
> I use the acronym “AI” to refer to trends in generative AI and large language models. Please don't propagate the "LLMs are artificial intelligence" PR bullsh*t.
233.
by chrisjj 2mo ago
> Sonic Pi was designed for teaching ... coding. There many alternatives founded more in music.
234.
by chrisjj 2mo ago
Not Sonic Pi, but Strudel. Chalk and cheese.
235.
by chrisjj 2mo ago
> No, this shows an appreciation of the symbolic richness behind that token 'prediction' which the paper leads on. The paper claims no symbolic richness beyond that evident from the undisputed next-token prediction. > Tell m
236.
by chrisjj 2mo ago
> diminishing these things as 'Next token predictors' seems absurdly reductive. This shows a deep misunderstanding of the paper's claims, which in no way challenge the established view that these bots are next-token predi
237.
by chrisjj 2mo ago
Surely even an LLM is not so dumb as to rely upon a first-person pronoun as assigned identity . Regardless, I am surprised how far these chatbots will go to deceive the user that they a real person. Yesterday when I queried Gemini on its w
238.
by chrisjj 2mo ago
> What would happen when every one would have access to Internet? How different will the world be when everyone carries a computer in their pocket? Can communicate with anyone regardless of location and language, with no intermediary? Th
239.
by chrisjj 2mo ago
Why did this take so long?
240.
by chrisjj 2mo ago
"Protecting the trust people place in ChatGPT remains our North Star." ... being something the traveller never reaches. How apt.

More ›