pj_mukh

AGI is undefined, AI is normal technology. Lots of academic works have analyzed this [1] and there is nothing, other than marketing hype, that supports this. It is "grown" is a meaningless term, because what do you even mean by that? Datasets are iteratively shaped? Grown is a very weird term for that.

AI may have continually extra degrees of freedom, but civilization only has so many modes of catastrophic failure. I don't grant the comparison but even nuclear technology has been massively useful and its main mode of catastrophic failure was brought under control via multi-national treatise. And I see no evidence that AI (outside of the marketing hype) is as dangerous as Nuclear technology.

[1]: https://knightcolumbia.org/content/ai-as-normal-technology

Yes so that’s someone designing a system (harness or prompt) to be dangerous. In all other systems we blame the designer not the system.

It’s like blaming Boeing for 9/11. Planes and AI are useful for a lot more than just terrorist acts. I have no doubt we’ll build a TSA for AI, and a lot of it will be security theater.

"What gives a paperclip maximizers purpose?"

The human who gave it the optimization function? That should seem obvious. If you take the biggest, best model in the world right now and put it in the box and give it no instruction it will do....nothing. I think you agree with that point, a lot of the hysterics right now is people not accepting that and it's useful to get on that common ground.

So given that most of the rest of the fear is around "let's not make scissors because some people will use them to stab people". Which is a fair argument and we probably do need to think about scissor safety but "ban scissors" doesn't quite flow from that.

Jesus that is an insane headline compared to what is actually happening.

>>ecological niche

As in..to be dominant? Why would an AI try to dominate? What would give it purpose, or is this a purpose via misalignment scenario?

I am trying to think through the scenarios here, like a biolab making something that a very advanced open source model prompted by some terrorists comes up with?

Why would a biolab capable of making something like be unregulated? And if it definitely would, isn't the problem with the biolab?

It feels like all these scenarios are leaving some gaping holes in our security infrastructure that have nothing to do with AI.

Misuse how exactly?

Can I ask you back, what your concern is here? It'll get so good so as to desire to hurt us or is it a misalignment event that you think will lead to disaster?

Or is it simply that you feel bad for Mathematicians.

Though the horsehoe is the horseshoe.

ALPR's have been a thing since the 1980's, nobody's been moving any windows, anywhere.

ALPR's pre-date Flock. The bill never mentions the brand "Flock". You could trivially mandate some stringent data retention policies, block cross-camera tracking if you really want, but just like his AI bill, a blanket ban is all this wing of the party can think of?

It's DOA, it'll never pass. He has the mandate to actually fix the privacy issue and he's p*ssing it away.

>> "which would prohibit federal agencies from using automatic license plate readers (ALPRs)"

...yup makes sense, no reason for federal governments to have unfettered access to this data.

>> "and block federal funding from state and local governments that use the cameras."

Wait..what? Why put this silly Poison Pill in the bill? So a city can't enforce its own traffic laws without having to put an over-zealous police force in the loop? The same guy that is handling drug dealers is stopping you for traffic tickets and people think nothing should go wrong?

Without cities adopting some massive road diet (fat chance) a reduction in ALPR's will kill people. Traffic deaths are at the same level as gun deaths in the USA.

The same one that is paying for their support for data centers and is shifting as a response [1]. Like I said, if political capital is being spent shifting peoples' stances, how much of a waste is it to know we could move them on this and we chose not to. Generational fumble.

[1]: https://thehill.com/homenews/senate/6121776-gop-worried-data...

Along those lines just a reminder folks, China is installing about 1GW equivalent of Solar power per day [1]. Lot of Australians can simply stop paying for power now due to free power on offer and batteries [2]. These are not zero-regulation, socialist wastelands.

All this whinging about "is this project using too much power or not" is pointless scarcity that we've chosen to partake in. If we are to create and spend political capital via hyper-focussing on a single issue, perhaps we should use it to fix this instead?

[1]: https://reneweconomy.com.au/just-staggering-china-installs-1...

[2]:https://www.energy.gov.au/rebates/solar-sharer-offer

I’m sure there will be a better term but that’s my term for a multi-modal searchable vector-store where you can use words to search spaces ala “that sculpture where the man is thinking on a door?” Which returns a physical point in an indoor or outdoor map.

Right so that gets you inflation adjusted pensions/social security, unlimited healthcare, even in the healthcare black hole that is America.

But that’s still not enough, you also need preferential rent control for wherever you want to live for the remaining few of boomers who don’t already own the largest swath of real-estate.

This now just looks like YIMBY’s screwing over old ladies out of their homes.

I empathize with a universal right to housing, but is that a universal right to a house in East village NY, Marina SF or El Retiro Madrid?

For everybody?

Or only for those who got in, in the 90’s?

How does that seem fair?

These details have to be worked out before you can get the public to commit to one of these and just rioting in response will only get you so far.

Oh boy.

I don’t want to freak the museum out but I have 360 degree footage of large sections of the museum (mostly for a spatial index) and now I’m wondering how high a quality I could get with splatting the sculptures and making it public.

Would I get cease desisted for hosting from across the pond? Or would they be too lazy to come after me?

Vibes. All headlines and vibes.

turns out consent, agency and autonomy are important!

to be clear, I suggested this for me to help me.

What does an LLM with zero instructions do nowadays with no loop? It's basically just noise spilling from the weights and "snapping" to the closest actions/tools , with no real signal going through?

Actually man freaks out and starts hitting the walls.

But either way, setting that asides by “give it full access to the internet” I meant giving it the tools to access the internet.

It just won’t though. It has no need to.

Another test I like for this: Put the LLM/agent/AGI/SI in a box with full access to the internet. Don't give it any instructions. Nothing. Not even a system prompt. What happens?

We know what the human mind does in that scenario, but even the most advanced unreleased frontier agent right now just sits there doing nothing. It doesn't really want to do anything. Seems like a crucial distinction to me.

I want a red light/green light for my Office Desk.

Red meaning "I'm working on something important. Disturb only if emergency"

Green being, "I'm scattershot, come hang"

This is the perfect platform for stuff like this.

ugh, want this so bad.

I bought so many sets, and I want to rebuild them, but they're a whole big mess. I want to sort them out.

Makes finding pieces for new builds much easier too

I know nothing about crypto, but wasn't one if its selling points that its "digital gold" for people without stable fiat currencies?

You send aid to a drought-stricken farmer to some sub-saharan region with ample cheap data but no (stable) banks, how...are they supposed to store the money?

I mean, the wars are also getting automated now. Sooo, no need to worry?

Occam's razor: Going to closed-source just to hide KV-cache optimizations seems silly?

Shoutout to people who don't overthink blog setups. "thisismypersonalblog.com" AND GO. It'll do the job.

On a technical note, your setup is not far from where professional setups are headed [1]. Astra is really doing an end-run around (for now) research Robotics setups.

[1]: https://x.com/ihorbeaver/status/2104646736652447854?s=20

"OpenAI models attack websites and take anything they can out of them because that is the business model of the company"

Wait, what did they hack and take into training data? On the whole, I totally buy that OAI is legally (and criminally) responsible for their agents, but surely, damages have to be proven?