@xjincaox I feel that pessimism as well as a new PhD student!
I try to take the opportunity to ask more fundamental questions + upskill on my weaknesses. while my research may eventually get undermined by AI, as long as I have the skills and intuition, I can always update my questions
I think this preprint title reveals a lot about Anthropic's approach to science.
Scientific paper titles basically never mention the tools used to make the discoveries they report. The focus is on what was discovered.
For example:
- "A Structure for Deoxyribose Nucleic Acid" (Watson & Crick)
- "Streptomycin, a Substance Exhibiting Antibiotic Activity against Gram-Positive and Gram-Negative Bacteria" (Waksman)
- "A Programmable Dual-RNA–Guided DNA Endonuclease in Adaptive Bacterial Immunity" (Doudna & Charpentier)
- "Design of a Novel Globular Protein Fold with Atomic-Level Accuracy" (Baker)
In each of these, the discovery / key deliverable is front and center. The machines, methods, and people used to make that discovery do not get mentioned because they're not the point. The discovery is what matters.
Yet, here, Anthropic chose to title their preprint "Autonomous AI agents discover reverse transcriptases with tandem repeat arrays."
That's not a scientific paper title: that is a headline.
It puts the discovery secondary to the fact that "Autonomous AI agents" made it. If they were serious about putting science first, the fact that AI agents found a new enzyme would be no more important than if Taylor Swift did.
This choice gives the game away. Though this is dressed-up as a scientific paper, it is not. It is a press release.
Mathematics has now fully morphed into biology:
Project costs millions $ ✅
Put out paper that nobody has read ✅
Massive paper supplement nobody will read ✅
Advertise with pretty art ✅
Vicious authorship fight ✅
Say you're curing cancer ✅
@chaitjo Antibodies are very hard to get right + model-as-a-service in this space means you need a really strong moat to stand out. Its just easier (esp as a startup) to leverage model performance as your moat, so you're disincentivized to make your neatest tricks public
Anatomy of BioML papers:
1. create benchmark misaligned w/ REAL biological goals
2. develop ML model w/ jargon-maxxing
3. compare vs models that are inappropriate or out of context
4. give it hypiest name — foundation model 3 years ago; virtual cell last year; world model now
What window? What technology in history has ever looked like this? This is the dumbest FUD ever conjured up in polycule group texts by a retarded cult in SF that thinks they will be the high priests of intelligence, ironically not only will they not, they will commoditize what gave them power to begin with. Midrange consumer hardware will train it in 20 years, but if you are motivated with a good sized hobby budget you'll be able to train it in 6.