@okTurtles @Group_Income @espionageapp #Fediverse preferred: https://xitter.cf/t.co/hYJJrmMskX To (maybe) see my posts on X: 🔔 or lists
Joined September 2008
- Tweets44.5K
- Following3K
- Followers3.4K
- Likes86K
. @deepseek_ai's latest v4.1 flash model may be interesting, but it is also scary.
I asked it to do something (through @AskVenice) and it just replaced my instructions with a total fabrication.
A reminder to keep LLMs on a short leash, and to scold anyone talking about "AGI".
Check out our new Astro-ified blogs.
They look identical to the Wordpress blogs, don't need updates, load faster, and way more secure:
taoeffect.com/blog/
fixingtao.com/
Tired of having to upgrade #Wordpress and #PHP?
Try a static site generator instead.
You can keep your comments.
We've successfully converted 3 blogs using this method to @astrodotbuild, just point your agent at:
blog.okturtles.org/2024/10/c…
Greg Slepak (@[email protected]) retweeted
Now is a good time to learn about DPKI:
okturtles.org/dpki
If you're wondering whether any of the 286 comments mentioned DPKI on HackerNews btw:
news.ycombinator.com/item?id…
Begging YouTube to reinstate your account when they inevitably ban you looks weak and reinforces a slave mindset.
Speak freely and move to a better platform.
YouTube has banned The Bitcoin Conference channel.
All our channel does is post keynotes and panels from the largest Bitcoin conference in the world, bringing Bitcoin education to the masses.
Please reinstate our channel.
@YouTube @TeamYouTube
DPKI fixes this btw.
okturtles.org/dpki
Companies do this thing where they spam you and then for you to login to unsubscribe.
Should result in immediate jail time.
. @deno_land will live on.
We will stick with it.
It is better designed than Node, and the software maintenance landscape has completely changed.
"Store your crypto in a device purposefully designed to store crypto", they said.
bleepingcomputer.com/news/se…
What I've realized: the reason some people look to authority figures in obviously corrupt institutions is because *they want someone to lie to them.*
They genuinely prefer bull💩 to the truth.
I am on the side of creationists and evolutionists!
And I am against creationists and evolutionists who are against one another.
Both creationism and evolution are true.
And you should both work together to discover the full truth.
I see a lot of people freaking out about Larry Ellison "owning all of the media".
A better question is why do you care?
Do you watch the media?
"Oh no this guy now owns my brain I am a slave to him I can't help not watch and obey his media!"
That's what they sound like.
Transformers are a threat to software innovation.
They put ecosystem pressure for software fossilization.
As useful as they are, the AI world needs to switch to something better that doesn't have this property.
We need AI that can learn new information on the fly.
Discourse is among the worst forum software.
I don't know why anyone uses it and it annoys me that so many new forums use it.
If you want to run a forum, choose Simple Machines Forum. It's the best forum I have run across, and I've been running two forums using it for years.
Exactly what I said.
Contrary to certain bogus comments on HN, programming is not "solved".
Transformers are fancy autocomplete.
None of this is surprising to anyone who bothered to study the architecture.
Womp womp.
Humans do this too btw, as with May's Theorem:
groupincome.org/2016/09/misc…
BREAKING: OpenAI’s solution to Navier–Stokes does not match its Lean verification.
The most important article to read today is not one of OpenAI’s 700 AI-generated math papers.
It is this other paper, making a deep and worrying point:
A Lean-verified proof does not automatically validate the proof written in natural language, nor does it mean that the formal statement captures the intended theorem.
During translation, an AI can change an assumption, weaken a statement, or replace the argument entirely.
It can hallucinate another theorem.
Lean correctly verifies the result.
But the proved result may no longer be what the paper claims.
This is a general problem. Things get spicy when the authors examine OpenAI’s proposed Navier–Stokes solution.
They identify at least two mismatches between the written intermediate results and their Lean counterparts:
One estimate claims that four additional input derivatives suffice. The Lean version requires five: a weaker result.
A pressure-flux estimate is obtained through a different bound, and proved through a different argument.
It is not clear whether these mismatches invalidate the entire proof.
But they raise an important issue.
OpenAI is flooding us with claimed revolutionary breakthroughs. Yet nobody knows whether the proofs are correct or whether they prove what they claim to be proving.
Epistemia at scale.
*
Paper in the first reply
