![Ollie Turnbull Profile](https://pbs.twimg.com/profile_images/1708487995051941888/dUejh_3K_x96.jpg)
Ollie Turnbull
@oliverturnbull1
Followers
170
Following
115
Statuses
157
Language models for antibody design @OPIGlets @UniofOxford
Oxford
Joined February 2011
RT @dwarkesh_sp: This question is even more puzzling and salient given the existence of Deep Research
0
263
0
@kareem_carr Reminds me of @DavidDeutschOxf arguments against physical basis of probability & proposing constructor theory
0
0
0
@DanHendrycks @ai_risk @scaleai Why make the benchmark public?? I worry that this will very quickly leak out and make it hard to evaluate real improvements
0
0
2
@mbeisen One thing is looking at the failure modes of large predictive models. Where these models break down might give insights into biological systems - showing gaps in our understanding where conceptual frameworks need updating and/or data collection needs to be focussed.
0
0
2
RT @king__choo: why i think multimodal long context is (still) underrated in the last 18 months, debates around long context vs rag (retri…
0
1
0
@nikosmela @nickcammarata @catehall Buddhism it just referred to meditation. In Buddhism refers to 8 distinct altered states of consciousness, brought on by high states of conciseness (or collectedness) of mind
0
0
0
RT @Ashwinreads: i don't know when we'll have agi. it could be 6 months or 6 years. but what i do know is that 43 years ago krishnamurti pr…
0
62
0
@nayafia @jhanatech Nice report - did you look at % accessing jhanas stratified by online vs in-person?
0
0
0
@liambai21 @etowah0 Very cool, biggest challenge here vs NLP is being able to automate feature labelling, do you have any thoughts on this?
1
0
1
RT @OPIGlets: OPIG DPhil student @oliverturnbull1 and postdocs @AlissaHummer & @mijr12 contributed computational profiling to a comparative…
0
7
0
@nickcammarata @imitationlearn Ok but what about tasks where you need to go significantly ood like creative hypothesis generation, do you think current paradigm will also get there? Seems very different to reasoning within domain has been trained on.
2
0
3
RT @OPIGlets: DPhil student @oliverturnbull1 has built pIg-Gen, a language model for paired-chain antibody generation. It creates diverse,…
0
22
0
@nickcammarata Thanks! Yeah I’ve found quite a bit of progress practicing jhana from rob burbea’s retreat
0
0
2