@nicbarkeragain These mental models are the actual source code. And they vary wildly from person to person. Feynman for instance discovered that Tukey modeled counting visually. And counting is just about the simplest algorithm there is! https://t.co/9eeJ7EuwaM
@__venki__ Can you say more about why “check out this technical trick” posts feel obviated?
Simon argues here that hoarding tricks is even more valuable now: you can ask a coding agent to compose them. (Hoarding and publishing aren’t the same thing, of course.)
https://t.co/kCo5pnVkq3
Voice Mode = Airhead Mode. Maybe the low latency requirements of realtime conversation do require a different model entirely. But another possibility is that we're all airheads in voice mode, and the smart, deep conversations are the written ones.
https://t.co/4H58h4MflU
Whatever the reason for this, it's unfortunate. I like the spoken dialog modality. And most humans probably prefer it. But reading back through a voice mode transcript, the lower quality of thought is unmistakable.
@Swizec Was also surprised by this! But I've used Huckleberry daily for 3 years now and love it. Has scaled well to 2 kids. Check out the CSV data export btw.
Also, congrats man!
@johnloeber Market for Lemons effect? I wonder how things compare to earlier eras. In the late 00s I did technical interviews en masse, and the low applicant quality also blew me away. Found solace in the following explanation.
https://t.co/cRcVFQqncX
• Our year numbering system is Christian
• Our month names are Roman
• Our day of week names are Norse
• Our 7 day week system is Babylonian
• Our 24 hour days are Egyptian
• Our 60 minute hours are Persian
• Our 60 second minutes, Persian
• Our use of base 60 is Sumerian
@Noahpinion Assume the set of interesting math problems is finite, i.e. there are N of them. But problem N+1 is interesting because it's the first problem not in the set. Therefore, the set must be infinite.