gradenko_2000 posted:

I feel like I'm maybe too dumb to elucidate this point well, but I've always had this feeling that the advent on computerization has created this paradigm where a lot of jobs end up becoming more about wrestling with the specific computer program that you need to document or interface with the digital accoutrements of the job, rather than doing the job itself.

Or rather, wrestling with the computer is (or should be) a job all on its own, that used to be handled by secretaries and clerks, but those got eliminated (because of capitalism's broad desires to do so) and it got passed on to people who might otherwise have something better to do.

Javid posted:

someone train that AI on pictures of cleaner bongs, good lord

Second Hand Meat Mouth posted:

how is a dog supposed to clean a bong

i say swears online posted:

isopropyl 90%

i say swears online posted:

isopropyl 90%

FFT posted:

Don't forget the salt to make the process a bit faster.

quote:

With the help of artificial intelligence, Chinese scientists have found a new weight-loss drug. And it can also treat type 2 diabetes.

The medicine, developed by AI-powered drug discovery company MindRank, has now entered its phase 2 clinical trials.
According to the Hangzhou-based company, the drug should also be able to hit shelves much faster than usual.

Called MDR-001, it works by binding to the glucagon-like peptide-1 receptor (GLP-1-R), a well-established drug target. By binding to this receptor, the drug stimulates insulin release from the pancreas, which helps to lower blood sugar levels and reduce weight.

quote:

Jin Xurui, an AI drug discovery scientist from MindRank, said: “The preclinical development of a pipeline takes three to four years to complete. However, MDR-001 obtained Investigational New Drug (IND) approval from both the FDA [Food and Drug Administration] and NMPA [National Medical Products Administration] in 19 months, effectively doubling the speed.”

This remarkable achievement was made possible by MindRank’s self-developed AI computing platform Molecule Pro and data centre, which contribute at every stage of drug development.

The team also collaborated with researchers from Westlake University and published an article explaining the inner logic of the model at the International Conference on Machine Learning (ICML) 2023, in February.

To start, a data-digging assistant constructs disease databases and automatically filters out potential targets for drug molecules to bind and take effect. Subsequently, a protein structure prediction calculator decodes the dynamic structure of target proteins, enabling the precision drug design module to create tailored drugs from scratch, providing multiple possible solutions.

crepeface posted:

does this thread include non-art AI posts?

https://twitter.com/SCMPNews/status/1707745710018380124?t=fVJEQ2gRW5HlPsZGUkSkIw&s=19

webcams for christ posted:

so every author whose research is in the training data will get credit & compensation, right?

webcams for christ posted:

so every author whose research is in the training data will get credit & compensation, right?

crepeface posted:

do you... do you think it's bad that useful drugs are being developed twice as quickly than they otherwise would

webcams for christ posted:

new medicine twice as fast good; proceeds from lifesaving drugs going to VC coffers bad

Zodium posted:

it's a chinese company.

Dokapon Findom posted:

Computers are so god drat smart

Echo Chamber posted:

https://x.com/Pioldes/status/1709323993847631978?s=20
No clue how they activated it.

Snowglobe of Doom posted:

i say swears online posted:

yeah but with salt [2007 voice] you can't evaporate the solution on a drying plate, scrape it up, and smoke it

Snowglobe of Doom posted:

I'm starting to see online ads for products which are clearly AI generated images.

The promo images:


The product photos:




I'm guessing they paid someone on Fiverr a measly sum to make those real world sculpts, hosed up hand anatomy and all

webcams for christ posted:

tbf I don't know anything about Chinese tax code, and how much redistribution is going on, but are Chinese Private Equity firms that do Series A VC funding necessarily Benevolent?

have a hard time sympathizing with firms that do MCM1 between patients and the drugs that can save them