Neuroscientist @__NERF | @BiologyKULeuven | @VIBLifeSciences | @leuvenbrainins1. Opinions my own. https://xitter.cf/t.co/VxTXlngcfo

Leuven, Belgium
Joined November 2016
A major problem with advanced LLMs is their ability to manipulate human judgment. Their outputs can fool users (and even their makers) into attributing consciousness and agency to algorithms, influencing how we think and make decisions.
1
1
95
College enrollments are not dropping because of incorrect assumptions about AI. They are dropping because the historical advantages of going to college outweigh less than less the costs of going to college. + foreign students are feeling less welcome in many countries.
Guy whose entire business model, which might be propping up the world economy, relies on an absence of widespread AI alarmism: "Being alarmist is bad" Online pundits: "Thank God for this brave and unbiased truthteller" losing mind, folkx
1
93
Bouke De Vries also has a project based on pro-natalist views... research.ugent.be/web/result…
23
What a terrible start for a new venture. "the rapid pace of innovation suggests we're approaching artificial intelligence" 🤣 even ChatGPT knows it's evidence free.
My journey to develop AGI spans 25 yrs, including 10+ yrs thinking about technical & societal perspectives at Google DeepMind. AGI is on the horizon - we need deeper understanding of its implications. To help, we've created the DeepMind Institute.
76
For example, these are the means of h-index across fields, which vary by a factor of almost 5. Does anyone believe that computer science has been less impactful than medicine? arxiv.org/pdf/1708.07893
This is terrible science. The h-index is crude and heavily biased, by no means a measure of excellence or merit. Group differences, even with statistical controls, can’t show that EDI caused “weaker” researchers to be hired. Canadian scientists should push back hard against this.
1
1
219
Replying to @DeblockBlock
Framing this as a left-right debate is misleading. Look, for example, at Cofnas’s most cited work. A high-school student could have refuted it. Like eugenics and “race science,” this is not truth-seeking, but an example of what happens when scientific standards are not upheld.
38
The project for which Nathan Cofnas was reportedly hired, “The Birth Problem of Liberalism”, casts religious minorities as a “challenge to liberal democracies” because of their birth rates. Can't imagine a more openly discriminatory and illiberal premise.
Nathan Cofnas' race science is bubbling up again so it's a good reminder that his entire research project is looking at contemporary social differences and then coming up with just so evolutinoary stories based on zero evidence whatsoever. He admits this! 1/n
2
3
330
Guess who paid for data centers and tax cuts for the wealthy in 2025? Taxes increased for 95% of the population. Six firms shared 40% of total tax breaks. 88 profitable companies paid 0% tax. Foreign investors got a larger share of the OBBBA tax cut than the lowest income 40%.
1
1
89
Replying to @DarioAmodei
Dario, you’re right: AI negativity isn’t about warnings. It’s about tech’s track record: privacy, exploitation, addictive products, deepfakes, hype and concentrated power. We appreciate Anthropic’s efforts to do better, but the results are mixed at best.
969
fixed it
major price cuts today: *80% drop for GPT-5.6 Luna, now $0.20 per million input tokens and $1.20 per million output *20% drop for GPT-5.6 Terra, to $2/$12 *GPT-5.6 Sol gets Fast mode in the API, up to 2.5x the speed for 2x the price, same intelligence
203