@rainbowjanelane@CovidSolidarit1 Ah yes , classic. I'm being treated for ADHD and of course they stuck a "health anxiety" diagnosis in my file because "l insist on masking" ๐.
@kafkaex It's a classic fallacy of treating exposure and "intrinsic attractiveness/dateability" as indepent variables but they are not . The latter will very often induce the former as people will gravitate towards you.
@anxiouscatmom True but even when it was "trendy" it was very disruptive and uncomfortable so they were only willing to keep up with it for a short while.
@ChristosArgyrop It has all been fortold ages ago by the eerily prescient song "ฮฮฏฮฝฮฑฮน ฮฯฮธฮฟฯ" by the great Dimitris Starovas : https://t.co/m9QWS9xQSM
@Phgenetix@sowmay_jain@laukiantonson Precisely. Given the inane reasoning mistakes that even today's SoTA models (think GPT-5.4-pro) make *all the time* I simply would not trust them to perform this level of expert analysis correctly.
@reactive_dude I built a primitive skill for each of them so they can directly pull messages from each other's session log with me and it's quite convenient ๐ซ
@MerrynSW@prothedogescam@business Well the idea is that whatever the 'next AI paradigm' is it will most likely run on GPUs, hence it is 'prudent' for the big providers to secure as much compute as they can, in advance.
https://t.co/5BhrXUR5lg
Highly courageous and commendable for a high profile person - in web dev circles at least - to wear a mask at a public conference in 2026! This is Simon Willison, creator of the Django framework.
@iruletheworldmo If "all they're doing" is scaling the same Transfomer based architecture from 2017 in all dimensions (pre/post training, inference time CoT etc) then colour me uninterested. Get back to me when there's a true breaththrough in architecture.
@deredleritt3r Show me the next genuine leap in architecture and I'll consider all that. For now it seems that we are stuck with an enormously scaled version of an idea from 2018.