is there any sign yet of the much heralded counteroffensive by ukraine? for the west and ukraine, if they were to go too early, ahead of all that hardware being in place, then we could see the front lines we see now in place for some time to come. after the g7 summit in hiroshima is a chinese attack on taiwan more likely likely or less likely? i think for now xijinping has made the calculation that he simply can t afford to invade taiwan. and should we lose sleep at night for fear that artificial intelligence, ai, is going to take over the world? we are going to have to have systems in place to police it. we almost need a new police force entirely focused on al. the city of bakhmut in ukraine now seems to be pretty much entirely in russian hands. capturing it took almost a year and cost thousands of casualties and all for a place that, right from the start, military experts said didn t have any discernible strategic value. but there is no doubt that belatedly the russians hav
of downing street earlier this week, has been released while under investigation. he was detained on thursday on suspicion of criminal damage and dangerous driving. hello and thanks forjoining me here at the bbc s headquarters in central london. is there any sign yet of the much heralded counteroffensive by ukraine? for the west and ukraine if they were to go too early, ahead of all that hardware being in place, then we could see the front lines we see now in place for some time to come. after the g7 summit in hiroshima is a chinese attack on taiwan any more likely likely? more likely or less likely? i think for now xijinping has made the calculation that he simply can t afford to invade taiwan. and should we lose sleep at night for fear that artificial intelligence, ai, is going to take over the world? we are going to have to have systems in place to police it. we almost need a new police force entirely focused on al. the city of bakhmut in ukraine now seems to be pretty m
version of it can pass the bar exam that s taken by trainee lawyers. now, obviously, it s not been six months in development, it s been several years in development. but it can write speeches, it can write computer code, it can do your homework. you know, there s endless uses for this one thing alone. and we haven t really seen this before. traditionally, ai has been good at doing one thing really well, but it s not been good at multitasking, so you d need a host of things to do all the differentjobs that you might yourself do in a day, right? but now this, agi it s called, general ai, is able to do lots of different tasks, you know, one thing after another very, very quickly. and it is very disruptive. what s the real danger, that ai will start to replicate itself way beyond any possibility of our control and then start bossing us around? so ai currently doesn t think for itself, you give it something to do and it does it. you know, taking the chat bot again as an example, it s got an
all the differentjobs that you might yourself do in a day, right? but now this, agi it s called, general ai, is able to do lots of different tasks, you know, one thing after another very, very quickly. and it is very disruptive. what s the real danger, that ai will start to replicate itself way beyond any possibility of our control and then start bossing us around? so ai currently doesn t think for itself, you give it something to do and it does it. you know, taking the chat bot again as an example, it s got an enormous database of knowledge. that database of knowledge is the internet, which is vast and as we know, full of rubbish as well. it s not necessarily the best source of knowledge, but that s what it s drawing on and it can do that very quickly and generally pretty well, but you have to keep an eye on it. but as it starts to be able to act more independently, if you like, to think about what you re asking it