@cahillbooks Who made this decision?. It really does need scrutiny. Dublin planning and preservation is an absolute disgrace. We can't just keep on with the apathy saying "ah sure... what can you do?".
@Johnnyhooha@cahillbooks@RachelMoiselle not hugely complicated at all. And how expensive&complicated is it to knock it down & build something else there. Apart from the fact that it is one of the few buildings from that era with valuable and distinct design language. It is an absolute disgrace to destroy it.
@EBryceLee@BradSchoenfeld don't think he was ever convinced that damage should be persued for its own sake. It really is not straight forward. And due to the conclusion being "we really don't know how this thing works" there is an inherent confusion in the individual assertions.
@BradSchoenfeld@EBryceLee OK so neither mild nor severe damage should be desired attributes of a training routine focused on hypertrophy?. The former presumably due to not producing the key mechs and the latter because of secondary effects. So are they really 'key mechanisms'?
@nathanclark_ Man!. You Nailed It!. I'm simply aghast!. The best research team, coupled with corporate mayhem and disastrous execution. It makes me so upset as I've always been a fan.
@baraklaniado@GaryMarcus LLMs do not reason in the formal sense. They appear to reason. They can solve problems which look like they are reasoning about the problem (when they are not). The harness around them does the reasoning & ability to encode a certain (finite) level of ontological representation.
@bindureddy AI haters are still there, you might not be looking in the right places. The folks screaming "it's just autocomplete!" were and still are absolutely 100% correct. Those saying it wouldn't be useful were just plain stupid. Ur statement is inept and unhelpful in the extreme.
@JamesOConnorTD I sincerely congratulate you on this. I really hope this is an earnest attempt to put things right and that we will have a proper science lead discussion leading to action. Establishing nuclear power in Ireland is actually the easy part, govt mismanagement will be the hard part.
@Mark_J_Henry By ur 'logic' we conclude that Irish ppl are in danger of malnutrition?. The most egregious thing about all ur posts is that u are normalizing a lack of critical thinking. If ur genuine, read some of the other responses to see how you can improve ur critical thinking ability.
@welovedonegal no question about it. Its Fintra, you can see largy rinn straighy ahead. Tide halfway in/out. Was afraid to spoil the fun and bias others, but it looks like its been up a while. What a place!. ;-)
@sulistuart@cormac_mcginley all i know is that they are very highly prized (if they are the same thing as 'percebes') in Galicia Spain. Very tasty. Collection off the rocks is a dangerous business, i think they generally stay below water line or like the splashing mix of water/air.
@JackAdlerAI@mana1951066@TheEuropeansHQ@DarioAmodei@AnthropicAI It could very easily be so different. And not just talent emigrates to SF, many of our best tech companies have sold to big US players rather than staying euro based (the local capital environment is a factor). Should be building our own platforms too. It is possible.
@AmandaAskell@gmiller its very much an averaging problem. but how to weight the average?. Reminds me of a quantum mechanical system wavefunction, for alignment we need a way to apply the right measurement so wavefunction collapse = the most representative answer(for all input states). Model = wave fn?
@AmandaAskell you're a unique point in a very high dimensional space, as we all are, but human's tribal 'engine' means projection on to an external left/right axis only allows binary categorical with arbitrary fuzzification!. Averaging seems like a necessary evil... i wish it weren't so.
@vanhalt@jarredsumner I was coming here to say the same thing. And in fact LLMs per se don't help as much as people think, and there are better technical approaches for many core issues, but in terms of disambiguated language understanding LLMs solve one of the biggest issues.