Thomas Hope retweeted
Women of #Iran tell their story of bravery, defying oppression and fighting for a most basic right. It’s simple yet powerful:
Thomas Hope retweeted
'#WorldStrokeDay is a pertinent reminder that stroke must be well represented in the new 10-year health plan and that the Government must engage with patients, carers, and health & social care professionals [to] inform policy decisions.'
Thank you @SarahDykeLD & @Uma_Kumaran. 👇
Thomas Hope retweeted
People have been rewriting history and saying that "everyone has always believed that LLMs alone wouldn't be AGI and that extensive scaffolding around them would be necessary". No, throughout most of 2023 (the "sparks of AGI" era) the mainstream bay area belief was that LLMs were *already* AGI, and that merely scaling their parameter count and training data size by ~2 OOM without changing anything else would lead to super-intelligence.
Might be time to give up on @wizzair. Advertised flight times often seem attractive. Actual flight times rarely are.
Thomas Hope retweeted
The iTalkBetter clinical trial is out: tinyurl.com/5up6cc7m It's a self-supporting naming app that helps people with aphasia make large gains in their ability to retrieve trained words. Connected speech also improved and iTalkBetter use caused changes in brain volume #aphasia
Testing the Disconnectome Symptom Discoverer model on out-of-sample post-stroke language outcomes.
It works! Despite there being quite a few differences between our data and what the model was trained to predict.
academic.oup.com/brain/artic…
Examining the surprisingly low-dimensional latent spaces identified from post stroke impairment data, via PCA.
TL;DR: these results don't necessarily imply that post stroke impairments are really low-dimensional.
academic.oup.com/brain/advan…
I am horrified beyond words at the atrocities committed in Israel. I genuinely cannot understand how anyone can imagine they might be rewarded for this, by any God worth worshipping.
Thomas Hope retweeted
The public in North America and the EU (not the rest of the world) is already scared enough about AI, even without mentioning the specter of existential risk.
As you know, the opinion of the *vast* majority of AI scientists and engineers (me included) is that the whole debate around existential risk is wildly overblown and highly premature.
Anti-vax sentiment is both bonkers and fascinating. Can anyone recommend any good, academic science on the topic? I suppose I'm assuming something medical / neuro / psychology / sociology, but happy to be surprised?
On the one hand, the history of AI is replete with claimed 'limits' that melt away as technology improves. But even accepting that, I found this article convincing.
Building artificial intelligence: Reward is not enough bdtechtalks.com/2021/07/07/a… via @bdtechtalks
New paper from Howard Bowman et al (incl. me) dissecting the confounds in studies of proportional recovery after stroke.
ahajournals.org/doi/abs/10.1…
Thomas Hope retweeted
Lesion site and therapy time predict responses to a therapy for anomia after stroke: a prognostic model development study biorxiv.org/cgi/content/shor… #biorxiv_neursci
Fascinating documentary series, this. Was a privilege to be (just a little) involved.
bbc.co.uk/programmes/m000mtm…
Thomas Hope retweeted
Moving beyond the dual stream account of language. New scientific commentary by Emily Upton & Thomas Hope bit.ly/3gu4lH8
One of the best articulations of modern political discourse that I've read.
Bringing proportional recovery into proportion: Bayesian modelling of post-stroke motor impairment academic.oup.com/brain/artic…
