Roberto Rossi (@gwr3n) 's Twitter Profile
Roberto Rossi

@gwr3n

BEng, MEng, PhD, FHEA โ€ข Full Professor & Chair in Uncertainty Modelling at @EdinburghUni โ€ข Head of Man. Sci. & Bus. Econ. @uoebusiness โ€ข #AI #orms #uncertainty

ID: 3131372524

linkhttps://gwr3n.github.io/ calendar_today02-04-2015 15:00:41

10,10K Tweet

324 Followers

400 Following

Subba Reddy (@postpcera) 's Twitter Profile Photo

Dmitrii Kovanikov disagree. If you want zero hallucinations, you had it for decades that is Database. you store whatever you want and ask for it (at atomic level) exactly the same thing with SQL. Hallucinations Are a Feature, Not a Bug sublayer.com/blog/posts/halโ€ฆ

Mehrdad Farajtabar (@mfarajtabar) 's Twitter Profile Photo

1/ Can Large Language Models (LLMs) truly reason? Or are they just sophisticated pattern matchers? In our latest preprint, we explore this key question through a large-scale study of both open-source like Llama, Phi, Gemma, and Mistral and leading closed models, including the

1/ Can Large Language Models (LLMs) truly reason? Or are they just sophisticated pattern matchers? In our latest preprint, we explore this key question through a large-scale study of both open-source like Llama, Phi, Gemma, and Mistral and leading closed models, including the
Toby Walsh (@tobywalsh) 's Twitter Profile Photo

"Overall, we found no evidence of formal reasoning in language models including the recent #OpenAI #GPT-4o and #o1-series. Their behavior is better explained by sophisticated pattern matching"

Prof Alice Roberts๐Ÿ’™ (@thealiceroberts) 's Twitter Profile Photo

I know people are leaving Twitter, but the algorithms also seem to be mysteriously (or not so mysteriously) working against some of us. I know that some followers of mine have somehow automatically unfollowed me (without them doing that), and that my posts are hidden or buried. I

ร–zgรผr Akgรผn (@ozgurakgun) 's Twitter Profile Photo

People posting LLM generated code as if they are correct is a great illustration of the danger of using LLMs, especially for reasoning tasks. Lower your guard and they will play you for a fool.

wakelyn (@theyfabatbirth) 's Twitter Profile Photo

my mom (who's a university professor) did something interesting last year: she assigned her students to give chatgpt an essay question, have it write a paper, and then proofread/fact check it. nearly every single student in that class came out of that assignment anti-chatgpt.

Jeremy Corbyn (@jeremycorbyn) 's Twitter Profile Photo

Israel has spent the past year desecrating international law on live TV. Now it is waging war on the UN itself, including 10,000 peacekeepers from 50 countries stationed in Lebanon. What will it take for the UK to end its military, economic & diplomatic support for Israel?

Jaclyn A. Siegel, PhD (@jacasiegel) 's Twitter Profile Photo

Nothing worse than going to submit a letter of recommendation only to learn that there are 10 ranking questions and 20 specific text boxes to fill out. Please, please, please just let me upload a letter of recommendation. I am begging you. No one has time for this.

Lord Bebo (@mylordbebo) 's Twitter Profile Photo

๐Ÿ‡ต๐Ÿ‡ธ๐Ÿ‡ฎ๐Ÿ‡ฑ In case it was unclear: 1) This is the yard of the Al Aqsa hospital 2) These are patience connected to life supporting medical machines 3) Theyโ€™re burning alive and canโ€™t get away

๐Ÿ‡ต๐Ÿ‡ธ๐Ÿ‡ฎ๐Ÿ‡ฑ In case it was unclear:

1) This is the yard of the Al Aqsa hospital
2) These are patience connected to life supporting medical machines
3) Theyโ€™re burning alive and canโ€™t get away
Owen Jones (@owenjonesjourno) 's Twitter Profile Photo

Iran's attack on Israel was a) aimed at military targets and b) didn't kill any Israeli civilians. The result? The UK imposing sweeping sanctions on Iranian officials and organisations. Israel wipes Gaza from the face of the earth. The result? The UK keeps arming Israel.

Grady Booch (@grady_booch) 's Twitter Profile Photo

For most of the history of computing, the correctness of calculations was essential: business use cases demanded it as did all of the embedded uses whose failure might yield considerable human or economic cost. At the turn of the millennium, the subtle turning of political

Elizabeth Laraki (@elizlaraki) 's Twitter Profile Photo

I'm talking at a conference later this year (on UX+AI). I just saw an ad for the conference with my photo and was like, wait, that doesn't look right. Is my bra showing in my profile pic and I've never noticed...? That's weird. I open my original photo. No bra showing. I put

I'm talking at a conference later this year (on UX+AI).

I just saw an ad for the conference with my photo and was like, wait, that doesn't look right.

Is my bra showing in my profile pic and I've never noticed...? That's weird.

I open my original photo.
No bra showing.

I put
Ryan Briggs (@ryancbriggs) 's Twitter Profile Photo

When I was young I fixed my parentsโ€™ computer and now that Iโ€™m older I fix computers for my kids. Are we the only generation that knows how computers work?

Ethan Mollick (@emollick) 's Twitter Profile Photo

It is morally wrong to use AI detectors when they produce false positives that smear students in ways that hurt them and where they can never prove their innocence. Do not use them. bloomberg.com/news/features/โ€ฆ

It is morally wrong to use AI detectors when they produce false positives that smear students in ways that hurt them and where they can never prove their innocence.

Do not use them. bloomberg.com/news/features/โ€ฆ
MaximumEffort433 (@maximeffort433) 's Twitter Profile Photo

Ryan Briggs We grew up on Windows 95, an operating system that just about wore its organs on the outside. Kids today grew up on tablets and phones with operating systems that barely let you get to root. 1957 Chevy vs 2024 Prius.

Reid Southen (@rahll) 's Twitter Profile Photo

As people keep saying, ChatGPT is impressive until you start using it for things you're knowledgeable about or an expert in, then you realize just how shoddy and unreliable it can be. Confidently wrong is the name of the game.