> The only thing that matters is if LLMs with sufficient scaling can become frontier AI researchers kicking off the exponential.
I think we know the answer to that already - LLMs show no sign of improving intelligence and instead providers are going down the ‘agentic’ rabbit hole.
There are too many things missing, like a world model, understanding, and taste (in the sense of knowing what is good and what is not good).
These are word generators, not agents, I’m really not sure why people think they could be capable agents (ie independent) when they consistently ignore instructions, generate the wrong things and then double down when questioned, etc etc.
You’ve been sold something that simply doesn’t work for the purported use case (intelligence) and instead is like a stupid database of all world knowledge with the appearance of intelligence.
Useful tools at times (if you bear in mind their limitations), but not close to intelligent, independent agents.
They were the ruling class of half of England for a significant period. That is not marginal. It influenced laws, culture and language, though later displaced by the Normans who also had a profound impact. Were the Normans also marginal as they did not have a huge genetic impact?
You're now arguing about something which isn't even in the proposal from the EU, it's a hypothetical from the article which attempts to draw a parallel with existing physical id (I guess to demonstrate to people that this already happens regularly with passports, ID cards, driving licenses etc).
That existing physical ID is government issued, has an issue no, usually has a photo on it, often other biometrics nowadays, and is definitely traceable!
Absolutely such a proposal cannot be fully privacy preserving, because at the very least it gives away your age, but also probably other identifiers.Just like your browser that you're using right now gives away a lot of information which advertisers use to fingerprint you like fonts etc. This sort of solution requires a central issuer/authority to know about the tokens, there's no way round that as there is no way round governments issuing passports if you want border control.
The goal is that it'll preserve much better privacy than uploading passports etc to random websites, which I think it will.
Untraceable is your invention and obviously absurd and unworkable.
No scheme is going to issue cards/tokens that can be traded between people, that's obviously unworkable.
Untraceable is for the consumers of the cards obviously, but in order to have trust they need to be traceable for somebody to verify them. The government can already determine who you are and already requires you to prove that at certain points, this is not a new invention of government it is required for governments to function.
Agreed, and often the root cause is something else in a person’s life (stress, feelings of losing control etc ).
Adding new labels related to food doesn’t do anything to fix the underlying conditions and if anything distracts from them and points the blame at food and physiology, which is not often the root cause of over or under eating.
The tokens could be tied to the device and Apple account by a provider like Apple, in fact you don’t need to issue tokens, only provide a web api that Apple and other browser providers support, which attests age.
This is certainly something that can be solved technically if we want.
Yes I almost commented on that too, I am astonished that someone said ‘nothing of interest’.
That sort of adaptation makes it incredibly accessible and I love that they kept the original words yet somehow it all made perfect sense. I think Shakespeare works best in performance.
If it’s taught well I think it can be very entertaining. There are lots of levels to Shakespeare and lots of very basic comedy. Watchman in Macbeth etc. the motives of characters are also explained well.
> The thing is, a profitable company that sees an obvious efficiency staring it in the face is still going to take that efficiency.
Judging by this CEO’s vapid post stuffed with meaningless LLMisms, and the condition of this company, the efficiency savings don’t seem to be there and are at best illusory.
Good luck to any companies who think they’re improving operations by jamming generative AI (or worse unreliable ‘agents’ based on the purported intelligence of GAI) into all sorts of processes where they don’t belong.
We’ll see over the next few years whether the 10x efficiencies are real or a mirage.
I think we know the answer to that already - LLMs show no sign of improving intelligence and instead providers are going down the ‘agentic’ rabbit hole.
There are too many things missing, like a world model, understanding, and taste (in the sense of knowing what is good and what is not good).