harryprayiv @ harryprayiv @infosec.pub

Posts

0

Comments

38

Joined

10 mo. ago

During the US Presidential election: Yes

Now? Not at all.

IMO, they thought they were helping keep the narrative about Kamala positive (and before her, they censored criticism of a man so clearly demented and being shoved down people’s throats that it was laughable) but in reality they were just enraging their allies, making them into lifelong enemies. You can never come back from giving up your integrity and dignity to genocidal corporatist demagogues no matter how bad Trump is.

Greed stems from fear if you really stop to think about it.

Those rules actually seemed pretty reasonable to me. Shitty, obvious uses of AI will obviously not win Oscars.

Ahh the great hivemind pendulum where if something is hyped and has some notable cons, it’s time to throw it away completely and resist it without prejudice.

This is happening with crypto too. To me, it just signals that a person is highly receptive to adopting whatever the hivemind tells them to think without question.

Obviously crypto and AI have some properties that make them utterly horrible (especially in the hands of bad people). But they also have some properties that have the capability to revolutionize or accomplish certain things like no other technology can.

No one seems to acknowledge this dichotomy when they’re unflinchingly under the influence of the hivemind. For example, I’m 100% positive that this comment will get downvoted heavily pretty sure I’d be ratioed for saying this on Mastodon.

Utterly backwards.

This thinking actually leads to Elon Musk worship and the dehumanization (and eventual massacre) of poor people.

This is what the most depraved sociopaths believe. They constantly repeat this lie to themselves to excuse themselves for utterly monstrous deeds.

Almost every horrific thing that humans engage in stems from fear.

Better, more fair voting systems: https://www.starvoting.org/

Better programming languages beyond even Haskell and Rust: https://www.unison-lang.org/ or https://granule-project.github.io/granule.html

better social networks (the one we’re on)

better systems of government

IIRC, SixOS is built for MIPS systems and runs quite well on them with lots of VERY futuristic options (the most impressive he calls “ownerboot”).

https://media.ccc.de/v/38c3-sixos-a-nix-os-without-systemd

https://codeberg.org/amjoseph/sixos

Live, laugh, love

*their

Modern day politics is about as staged as professional wrestling.

Stop being as gullible as a Hulkamaniac.

https://youtu.be/tQhWXZvqwpM

Another suggestion I have yet to see mentioned is Zig. It’s easier to learn than Rust but just as powerful. It’s arguably more ambitious than even Rust because they are approaching LLVM as well as building a C/C++ replacement.

As long as he doesn’t try to make peace with Communist nations, our MIC won’t kill him like they did to JFK*.

• That’s now a fact since the Kennedy files have been released.

Purescript is also really great if you want what Typescript promised you but failed to deliver.

Haskell is my main driver and I think Haskell’s error messages may be even better.

Here’s another technology aimed at a similar idea: https://privtracker.com/

John Cazale

🙏

To understand what's actually happening, Anthropic's researchers developed a new technique, called circuit tracing, to track the decision-making processes inside a large language model step-by-step. They then applied it to their own Claude 3.5 Haiku LLM.

Anthropic says its approach was inspired by the brain scanning techniques used in neuroscience and can identify components of the model that are active at different times. In other words, it's a little like a brain scanner spotting which parts of the brain are firing during a cognitive process.

This is why LLMs are so patchy at math. (Image credit: Anthropic)

Anthropic made lots of intriguing discoveries using this approach, not least of which is why LLMs are so terrible at basic mathematics. "Ask Claude to add 36 and 59 and the model will go through a series of odd steps, including first adding a selection of approximate values (add 40ish and 60ish, add 57ish and 36ish). Towards the end of its process, it comes up with the value 92ish. Meanwhile, another sequence of steps focuses on the last digits, 6 and 9, and determines that the answer must end in a 5. Putting that together with 92ish gives the correct answer of 95," the MIT article explains.

But here's the really funky bit. If you ask Claude how it got the correct answer of 95, it will apparently tell you, "I added the ones (6+9=15), carried the 1, then added the 10s (3+5+1=9), resulting in 95." But that actually only reflects common answers in its training data as to how the sum might be completed, as opposed to what it actually did.

In other words, not only does the model use a very, very odd method to do the maths, you can't trust its explanations as to what it has just done. That's significant and shows that model outputs can not be relied upon when designing guardrails for AI. Their internal workings need to be understood, too.

Another very surprising outcome of the research is the discovery that these LLMs do not, as is widely assumed, operate by merely predicting the next word. By tracing how Claude generated rhyming couplets, Anthropic found that it chose the rhyming word at the end of verses first, then filled in the rest of the line.

"The planning thing in poems blew me away," says Batson. "Instead of at the very last minute trying to make the rhyme make sense, it knows where it’s going."

Anthropic discovered that their Claude LLM didn't just predict the next word. (Image credit: Anthropic)

Anthropic also found, among other things, that Claude "sometimes thinks in a conceptual space that is shared between languages, suggesting it has a kind of universal 'language of thought'."

Anywho, there's apparently a long way to go with this research. According to Anthropic, "it currently takes a few hours of human effort to understand the circuits we see, even on prompts with only tens of words." And the research doesn't explain how the structures inside LLMs are formed in the first place.

But it has shone a light on at least some parts of how these oddly mysterious AI beings—which we have created but don't understand—actually work. And that has to be a good thing.