The point is that the referendum was an unconditional 2015 Conservative manifesto promise, and was going to happen anyway, regardless of how Cameron's renegotiations with the EU went (sources here: https://news.ycombinator.com/item?id=49726666)
The outcome of the negotiations may have affected the course of the Leave and Remain campaigns, but it was not the trigger for the referendum itself.
>[The Conservatives promised a referendum] iff the renegotiations with Brussels failed, which they did.
No. The Conservative party promised an EU membership referendum by 2017 in its 2015 manifesto. That was not conditional on a failure of renegotiations. Nor, indeed, did David Cameron even concede that the renegotiations had failed, so that can’t possibly have been his stated reason for calling the referendum.
> My policy is to hold a renegotiation and then a referendum. That is what we promised in the manifesto and then to abide by what the British public say.
So you can see that there was never any suggestion that the referendum would be called or not called depending on the outcome of the renegotiations.
No, nothing public, we are all speculating. But we can surely guess at 99% that Canada would get no voting power, being that we are in North America and Canada isn't interested in that kind of arrangement.
Humans don't need a lot of training and finite tuning to make only legal moves.
An intelligent adult could simply read a short summary of the rules of chess and then, if they were careful, play a very bad game of chess without making illegal moves.
An LLM that has not been trained on any chess data cannot do that, at present. If you doubt it, take a current model and tell it that you want to play it at a variant of chess where, say, knights can also move diagonally like bishops. A human can easily adapt to this new ruleset (even if they make tactical mistakes, not having practiced with this variant of the rules).
How long a prompt do you think would be required to cajole an LLM into making legal moves at the rate of a human? Or do you think no amount of prompting could do that?
I don't know. My understanding is that current models will eventually fall into making illegal moves in longer chess games, and that no amount of prompting reliably gets them to stop doing so.
More importantly, beginner human players don't exhibit that tendency. The history of the position doesn't bother a human (except as required for castling and en passant rules), and the analysis becomes generally easier as pieces come off the board.
Nothing is forcing the LLM to play 'blind'. If it's smart, it should be able to create its own representation of the chess board and update it with every move, just like a human would. Any chess engine that's sensitive to how the moves are formatted is clearly not very capable.
A human wouldn't do that, they'd look at the board. I'm not disagreeing that to demonstrate clear superhuman ability the LLM should be able to do this, but it plays better than most humans blindfolded, and with fair prompts seems very good otherwise.
That's what a human will do if they already have a physical board to look at. But if someone, say, posed you a chess exam question via FEN notation, or as a sequence of moves in algebraic notation, you'd sketch a visual representation of the board off your own initiative to help you answer the question. There is nothing in principle to stop the LLM creating its own board representations in whatever format enables it to easily keep track of game state and legal and illegal moves. If it fails to do so, that's a sign of its own limited understanding of chess as compared to a human.
The LLM would only be playing 'blindfolded' if you somehow forbade it from making notes (as you effectively do by literally blindfolding a human, given how limited human working memory is). But you are not doing that. The LLM is free to keep track of the game state via whatever means it chooses.
None of this is about superhuman ability. Any human who understands a given chess notation can convert it to a visual representation of a chess board and then use that representation to choose their next move, with their usual level of performance.
I maintain that the amount of effort to teach a human to do this vastly outweighs the amount of effort to teach an LLM to do this unless you're deliberately trying to make them fail. I honestly have no bigger point than that, I just think this isn't a very good thing by which to evaluate LLM capabilities. If there's no argument you'll accept, I am happy to move on.
You don’t need to teach a human anything except the rules of chess and the details of a particular chess notation. No special skill or training is required to make a sketch of a chess board. Surely there is no chess player who, if confronted with a sequence of chess moves in algebraic notation, would not think to construct a representation of the chess board in order to understand what was going on.
> I just think this isn't a very good thing by which to evaluate LLM capabilities
I don’t think any single task is a good way to evaluate LLM capabilities, but I don’t see why chess is worse than a lot of other tasks. (Of course it is of no practical consequence whether LLMs can play chess, so if you are just making that point, then yes, I agree.)
> If there's no argument you'll accept
It’s a little unfair to suggest that I wouldn’t accept any argument whatever for your position just because I haven’t been convinced by your very brief comments so far. I could equally well say the same thing to you!
The issue with the models isn't that they play a bad game, but that they persist in making illegal moves. An average intelligent human can be told the rules of chess and then play chess, badly, within the rules.
Sure, but the LLM is free to construct a representation of the chess board and update it as it goes along. It is not in any way banned from using a virtual board, or whatever representation of game state it pleases.
AFAIK, current models will still sometimes make illegal moves even if given the entire game state (e.g. in FEN notation), so it is not purely an issue with the models’ ability to keep track of sequences of moves.
One use case I’ve found is for a lock that you don’t need to acquire. For example, you need a lock to read a cache entry, but if you can’t acquire the lock after a few spins, you can just proceed without the cache. For fine-grained locking, a spin lock can have a significantly lower memory overhead than a full futex.
I certainly wouldn’t go anywhere near Twitter these days, but the question is how much of that is to do with engineering failures and how much is to do with the overall change of culture and moderation practices that Musk has introduced.
Very much this. If you have 3000 engineers (and all the associated non-engineering roles that inevitably come with that) then you have guaranteed that no problem will ever be easy to solve. Possibly you might be able to solve some extremely difficult problems that you couldn’t solve with fewer engineers.
The iPhone 18 Pro main camera is f/1.48, which is half a stop wider than the f/1.78 main camera on the 17 Pro. An f/1.48 fixed aperture would probably not be ideal (less depth of field, more revealing of lens aberrations) so there is a real light gathering advantage to the variable aperture.
>Also a phone already starts with very deep depth of field
Fairly deep. You can easily see background blur on a modern iPhone camera if you focus on something close. The very fact that the cameras all have variable focus shows that depth of field isn't as deep as all that. The extra depth of field from a narrower aperture could certainly be useful in some cases, such as landscapes. Here's a question for your favorite LLM:
"If I am shooting an iPhone 17 Pro main camera, what aperture would I theoretically need to get perfect focus from 1.5 meters to infinity, assuming a CoC suitable for a 12MP shot?"
Answer: about f3.2 if you manage to focus at the optimum hyperfocal distance.
>which when stopped down physically reduces the amount of light from the scene that reaches the sensor
As you say, this is only the case if you can't increase the exposure time to compensate. In daylight you will often have plenty of latitude to do so, even without image stabilization. Typical daylight shutter speeds at f/1.78 can be around 1/1000 at base ISO, so there is plenty of room to increase exposure time by 2-3 stops in those conditions.
Come off it. Neither the referendum nor the referendum result were inevitabilities.
reply