Batteries not included.

It's trying to frame a philosophical statement mathematically. Or maybe the absurdity of trying to do so. The statement isn't rigorous by itself.

Ah, cheers. Was just wondering.

I think even % 2 = 0

EU AI Act going ahead

https://uk.pcmag.com/ai/151416/europe-passes-ai-act-to-ban-riskiest-tools-force-copyright-compliance

The AI Act's passage means that once the law is finalized and put into effect, AI systems will be categorized into one of four risk categories and be regulated accordingly. Any AI tools deemed "unacceptable" would be subject to an outright ban, such as biometric categorization, social scoring, emotion recognition, and predictive policing tools powered by AI.

When people encounter AI tools like chatbots online, the law will require companies to disclose to EU users that they're interacting with an AI, not a real person. "Humans should be made aware that they are interacting with a machine so they can take an informed decision to continue or step back," the commission says.

The European Parliament also notes that general-purpose AI systems (GPAI) will have to comply with EU copyright laws and publish "detailed summaries of the content used for training." Copyright holders are allowed to protect their work from text or data mining, meaning that as long as their work has been opted out from such uses, generative AI developers will require permission from the copyright holders to use their work.

Yeah. They say it's a first attempt at legislation. But no one knows what a second attempt is going to look like. Or how effective any of it is going to be. It's useful to establish some best practices and some no-go zones. Even if, in practice, there's a lot of bleed through into those parts.

The copyright stuff is necessary and urgent. Any commercial AI company that can't supply, on demand, every piece of training data they've used, should be shuttered. I don't know if a "detailed summary" will go quite that far. Probably still a lot of room for theft and lies.

Nvidia GTC keynote - blah blah GPUs blah blah AI blah blah exaflops and robots but OMG! skip to 14m:09s mark because a couple of the Disney little BD-1 style droids come on stage and it's amazing how much character the Imagineers and performers/controllers get out of the little robots (running on a Jetson apparently) - just blows away all these expensive humanoid style bots trying to badly drum or dab:



Disney knows what to do with robots (e.g. the animatronics and stuntronics) - all these startups trying to make human-shaped robots that can make a cup of tea but spill it everywhere and drop the saucers are pissing money down the drain. No one needs that. It's like the fucking Gundam mechs - "we need to fight the war by making giant fighting robots, with fully functional hands and then we'll build giant guns with triggers so they can shoot at the enemy!" Then the other country mass-produces flying guns at a production delta of about 1000-to-1 and tank-rushes to win the war. We need accurate solutions to specific problems, not mostly-inept general-purpose solutions looking for problems to solve.

Yeah, the little wobbly antennae really add to the charm. 
Some footage of them testing in the Park:





Robots designed for a specific purpose, rather than solutions looking for problems.

They'll come chump. The Jetson Nano is the point right now, and it's cheap. It means useful robotics is no longer in the hands of companies like Boston Dynamics. If you think of it like the Raspberry Pi of robotics you can see it'll get very specific use cases. It might look expensive but the core tech isn't, which is the entire point. You can just as easily swap the legs out with wheels and it'll figure out how to navigate.

Reinforcement learning like this is where AI is heading. It can go out and get its own datasets.

Here you go Chump, it's £140.







£500 for the new one, but that's equivalent to 80 old Nanos.

Well it was a stage show. They're not going to have flying robots tasering the audience.

Well they could I guess. Might bump up the share price a bit further.

LivDiv wrote:

Just don't expect people to be wowed by a robot doing a worse job if pouring a cup of tea than Michael J Fox.

You know that tech gets better right? In the case of AI it gets real good real quick.

LivDiv wrote:

It's the obsession with recreating human actions in these robots when there are better, simpler ways of performing those actions using robotics.

Probably said before it’s because we’ve shaped the work and home space and its tools to fit our form. A circular robot lawnmower can’t drive a forklift one day and fold clothes the next.

A general purpose robot means doing thousands of very minute and complicated things. You can show it how to boil a kettle and make a cup of tea. But it also has to open a cardboard box of tea bags with the stupid film thing around it. Can it find the little transparent film tab and remove it in such a way that only the top of the film comes off so the rest of the film stays in place around the box to catch the tea dust that always comes out of the bottom? Can it remove the perforated cardboard bit at the top of the box? Can it carry the flimsy bit of film over to the pedal bin? What happens when I change brand? What happens when the gf has overloaded the shelf we keep the tea bags on with all her weird fruity teas and it has to make executive decisions about what gets stacked where so the new box of tea bags fit? Can it root through the chaos of the snack cupboard to find the biscuits I want, remove the plastic clip we use to keep them sealed (of which we have three different types and it could be any of them), remove two biscuits without compromising the integrity of the rest of the wrapper and put the plastic clip back on? If it's a new pack can it find the little knife I use to slice the top of the pack, avoiding the plastic tab that the manufacturer's intentionally place way too far down the pack to encourage quicker consumption? We probably cycle through 20 different types of biscuit and each one needs the right decision. Can it place the knife in the safe place by the sink to avoid any accidents while it's awaiting wash up? Is the dog going to be sufficiently respectful of the robot's authority to not use the opportunity of the open cupboard to start snaffling around in there? Can it check the date of the milk? Open new milk? Another little tab that it has to grapple with. Move the rest of the stuff at the bottom of the fridge that's kept on top of the new milk, remove the milk, then place it back, ensuring that the more robust stuff stays at the bottom of the stack while the more fragile items are kept on top? Then rinse the old milk container out, crush it and squeeze the air out while simultaneously replacing the cap to ensure there's a vaccum and it stays flat then place it in the recycling?

The answer to all these questions is Yes. Eventually. And that 'eventually' might happen a lot sooner than we think, given the speed we normally expect technology to improve. But that's just tea. And I haven't seen anything that's even close to any of that. At the moment we're expected to be wowed that it can line up the water pouring from a kettle into a mug.