this is a pretty cool talk. it’s his first big big talk, and it’s a little all over the place. he’s very very against medical professionals, which is fair from his pov. im priviliged in that the ones ive had to interact with have been Fine/Positive. still not sure id compare ‘following doctors orders’ to following nazi’s orders during ww2, or that it’s a good argument in favour of taking agency over your own healthcare! healthcare is hard, there are experts in it. if they explain it to me, properly, then that’s good enough for me.
but the pharma industry is not that, and is much much worse.
i do like his shtichk of “here’s a lying pharma ceo, let’s call him. anyway, you can make drugs, here, i did it. you want it? there you go! keep that person healthy, keep that person safe”
“it’s not hard” mmmmm.
he’s very much like showboating anarchist propagandist lol. kinda abrasive but i guess that’s why he’s in the front.
very grateful for the nhs, and long may it remain!
oh huh, this dr. mixael laufer guy is a mathematician.
jesus christ pharma industry is ridiculous….. grrrrr!!
okay now on the how.
okay so their process has a few parts: - tool for working out the synthesis - tool for turning that synthesis into actual steps, recipe - tool for doing the chemistry
the former is clever computer stuff, the latter is a small flow-chemistry reactor. it’s quite involved tbh, but well within the capabilities of your local makerspace.
whilst he’s describing their machine, someone is assembling it live on stage. that’s pretty cool lol.
ugh this talk is actually two talks spliced together. it’s really long.
my tldr is wow this is really cool, i would have phrased/put things differently but like, can’t complain really.
walking through the hardware, he really likes push-pin connectors and screw terminals lol.
okay so it basically makes the drugs which are active pharmaceutical ingredient (api) + or two + solvent + filtration. which is a lot of drugs!! E and T for one…
they have divided the labour properly. human does the easy, human bits. computer does the annoying hard stuff like stirring at a certain rate for a long time.
woah, it can do some error correction like “are there still particulates?” keep stirring. branching and flow charts.
so big part is how simple they make it to make recipes. and how simple one-step synthesis pathways are. thank you chemists! your work is greatly appreciated!
okay so that chemhacktia thing is predicting reactions based on literature? not sure how it’s doing that. it is an open source tool developed by pro chemists, so i’d trust but it’s not like, 100% gonna work. how do you tell? hmmm okay he bsed that a bit. i give him being the salesperson, the group gets benefit of the doubt.
he makes a good point that all this technical know-how is only as good as its community. ah good he’s anti-llm. machine learning is probably different.
he’s got a weird glowing pda/pager at his hip lol