Well some of us are. Many top neuro journals straight up desk reject model organism papers nowadays. And if you want to get fast, detailed and precise mechanistic answers - some of these are still the best systems.
Very cool! Useful tool for #Drosophila folks. Would be great adopted by whole labs / institutions. Considering it's newness, add features? Tinkering & I don't yet see: 1) Retrospective dating / bulk stock flipping 2) Stock acquisition date 3) Column sort by chromosome 4) Reference usage citation
We had a spreadsheet too but when I imported it into this system I ended up throwing quite a few lines away because clearly we did not know what they really were, and could not trace them back. Things spiraled out of control at some point...
Happy to announce I've launched a new service for the fly community: flyRoom, a web-based comprehensive stock manager for all your fly pushing needs. Organise all your stocks, maps all your genotypes, plan your crosses, print and scan your labels. Give it a try and let your colleagues know!
Screen shot from the UKRI Funding Finder. Title: Fundamental AI Research Lab. Timeline Open 2nd March, intention to submit date 16 March, Closing Date 31 March 2026.
Can anyone think of an example of government spend of this magnitude and speed (excluding covid)??
4 weeks from call announcement to submission (no advance warning) and within that only 2 weeks to flag your intention to submit.
£40M total - individual awards £9.4M
www.ukri.org/opportunity/...
£9 million in AI is the textbook example of too little too late.
And all decided in 4 weeks? What is this? A joke?
An example of how useless X has become. 2020 Twitter would have been the place to go to get real time news on the war. Now it's 100% noise. Just old clips repackaged as new and an ocean of conspirationists talking to a crowd of blue checked clueless idiots. Incredible.
Richly perceptive analysis of UKRI’s funding ‘strategy’ by @ersatzben.com raises 2 questions: why do we still not understand how best to manage the interfaces between research & societal needs? And why doesn’t UKRI give Ben a job? open.substack.com/pub/johnsonb...
Someone has sure already made this observation but the fact they can convert all those empty warehouses into prison camps means they could have converted them into housing, community centers, job training centers or, hell, libraries or schools all along. It’s always a matter of will not resources.
There were 36 votes between the Green Party and Reform.
36 votes.
Reform are absolutely stoppable - and the Green Party can win with our message to lower bills, protect the NHS and rebuild public services.
Join.greenparty.org.uk
Labour’s vote is collapsing everywhere.
In Peterborough last night, as in Gorton and Denton, voting Green is the only way to stop Reform.
Screenshot of a home office tweet
If the PM reckons Jim Ratcliffe should apologise for spreading this exact same lie, should the Home Office apologise too, or? 🤨
A Reuters photographer captured this image of a page from Pam Bondi's "burn book," which she used to counter any questions from Democratic lawmakers during an unhinged hearing today.
It looks like the DOJ monitored members of Congress’s searches of the unredacted Epstein files.
Just wow.
It would be good to understand whether this deliberate castration of UK science and its international competitiveness is a vision of the government or of UKRI's directive.
So now we have it- not only the reduction from 12-13 to 3 grants from each board recommended for funding but the budget for applicant-led research to be reduced from £200 million to £113 million per year. This is appalling, Especially given all that is happening with medical research in the US.
That does read like a 1:34 am reflection indeed...
Paywalled but visible to many uni-based people
Apparently...
Much reduced MRC success rate for grants now in pipeline
Reduced MRC future funding for 'applicant-led' research
But 'life sciences remain a priority'....
Lot of room for more clarity (and recognition of effects on ECRs)
The media play along, but the real poison comes from those inflating the market bubble. Especially OpenAI. Anthropic has done better research and has stronger fundamentals IMO. I see confusion among academics too though.
Why should they be included? The point of an utilitarian definition is that is agnostic to these totally arbitrary inclusions.
That's a misunderstanding as old as Aristotles: all men are intelligent, all men eat pizza, therefore intelligence requires pizza. (Replace pizza with need to survive, etc)
That is fine. The point is not to demonstrate or claim that AI intelligence is fully overlapping with human intelligence. It is not and I don't expect it to be. (I kind of touch on that in section 6.5 when I talk about the strawberry test. There are important structural differences)
This is currently being done at multiple levels. Perhaps the most involved in this field is @yann-lecun.bsky.social who is pursuing JEPA as a way to separate "words" from "worlds". His position is interesting but part of his denial is ontological: may not pass my checklist! It's a fair gamble though
That's a step forward in the right direction :)
That goes all the way back to the "ape of genius" concession. I grant you something but I will conveniently change name to it :)
Interesting discussions on this topic (here and on other socials) made me realise that English provides an interesting peculiarity that other languages do not have. A distinction between intelligent and smart.
Q. for skeptics and everyone: would you say that LLMs are smart but not intelligent?
IMP 2002-2006. Did we overlap?
On the last point: intelligence is an observable trait which is why turing was wise enough to focus his imitation game on it. No introspection needed.
We can say "it's intelligent!" of a toddler or a dog. Still not clear to me why a machine should differ.
It's not, but the same is true for virtually all my trainees.
I don't think it's dishonesty either and I do agree that there is a component of that: "no battle plan survives contact with the enemy".
But there is an unscientific behaviour in moving the goal post that we do not condone in other circumstances.
Perhaps it's a sign we need to build better priori.
Sorry Dan- did you have a chance to read the paper? It's all in there.
I disagree it's hard (it depends on the precision) but I am not sure it's about measuring anyway because we are making a qualitative non quantitative argument.
As for refining the goalpost: after a while it becomes p-hacking!