LLMs do a cool parlour trick; all they do is predict “what should the next word be?” But they do it so convincingly that in the right circumstances they seem intelligent. But that’s all it is; a trick. It’s a cool trick, and it has utility, but it’s still just a trick.
All these people thinking that if only we add enough billions of parameters when the LLM is learning and add enough tokens of context, then eventually it’ll actually understand the code and make sensible decisions? These same people perhaps also believe if Penn and Teller cut enough ladies in half on stage they’ll eventually be great doctors.
If you’re 2x farther away, the intensity of the sound with be 1/4, because of the inverse square law, so logically your speaker would need to be 300% louder.
How do you think google maps got such a good database of these things? When I last used it, Google maps used to actively ask you for this sort of information.
My favourite is the “double tap with three fingers to zoom in the display”. This happened to me by accident while my phone is in my pocket. The first time it happened, I had to borrow someone else’s phone to look up how to undo it.
Another reason not to use EST/EDT is that they are overspecifying the time zone in most cases. If you ask for output in EDT, but the date is in December (which is not part of daylight savings time) then should the date output be UTC-4 or UTC-5? Technically you asked for EDT.
Using EST as a shortcut isn’t a good idea, either - most software will “know what you mean” and use EST or EDT appropriately... except, both The USA and Canada have been toying with the idea of dropping daylight savings time, so it’s very possible at some point in the future that 6:00pm on July 1st will be EDT in America/New York, but EST in America/Montreal (or vice versa). This is already true for CST - Saskatchewan doesn’t use DST. And there have been times where one country or another changes the start date or end date of DST too, so there’s no reason to assume those will always be the same between Canada and the US.
All these people thinking that if only we add enough billions of parameters when the LLM is learning and add enough tokens of context, then eventually it’ll actually understand the code and make sensible decisions? These same people perhaps also believe if Penn and Teller cut enough ladies in half on stage they’ll eventually be great doctors.