Adel Bibi

11.5K posts

Adel Bibi banner
Adel Bibi

Adel Bibi

@Adel_Bibi

Senior researcher in machine learning @UniofOxford. R&D Distinguished Advisor @SoftserveInc. JRF @KelloggOx. Ex-@Intel. @KAUST_News and @K_University alumnus.

Oxford, United Kingdom Katılım Ağustos 2010
1.3K Takip Edilen1.4K Takipçiler
Sabitlenmiş Tweet
Adel Bibi
Adel Bibi@Adel_Bibi·
Excellent news to share! I've been awarded the Systemic AI Safety Grant as one of 20 projects selected by the UK AI Security Institute 🎉 I’ll soon be hiring a postdoc to push the boundaries of Agentic OS Safety. Huge shoutout to my amazing co-PIs, @philiptorr and Adam Mahdi!
AI Security Institute@AISecurityInst

We've funded 20 new research projects to enhance AI security in critical infrastructure ⚡ Our Systemic AI Safety Grants Programme, announced at the Seoul AI Summit, has awarded up to £200,000 seed grants to projects tackling AI risks 🧵👇

English
6
0
25
4.2K
Adel Bibi retweetledi
erin griffith
erin griffith@eringriffith·
A detailed and brutal look at the tactics of buzzy AI compliance startup Delve "Delve built a machine designed to make clients complicit without their knowledge, to manufacture plausible deniability while producing exactly the opposite." substack.com/home/post/p-19…
English
203
405
4.7K
4.5M
Adel Bibi
Adel Bibi@Adel_Bibi·
great to see yann working against the flow again (one of a few people who do not believe llms are the path for intelligence) more importantly, YANN, with all his credentials, was offered 1B at ~4B valuation in the EU. let that sink in.
Yann LeCun@ylecun

Unveiling our new startup Advanced Machine Intelligence (AMI Labs). We just completed our seed round: $1.03B / 890M€, one the largest seeds ever, probably the largest for a European company. We're hiring! [the background image is the Veil Nebula - a picture I took from my backyard, most appropriate for an unveiling] More details here: techcrunch.com/2026/03/09/yan…

English
0
0
6
676
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen okey clear; you are a money driven lad who has not heard of scientific research before (unless you call that a product too) i expected more from this conversation, unfortunately
English
0
0
0
17
Austen Allred
Austen Allred@Austen·
“The guy will position himself anywhere there’s money to be made” and it’s a guy who has been quietly working on the same startup for a decade
Austen Allred tweet media
English
42
78
1.9K
445.8K
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen do we learn coding to do products? not sure whats ur definition of a product is. because it could be so abstract that encompass everything
English
1
0
0
51
Austen Allred
Austen Allred@Austen·
@Adel_Bibi They helped people build products before, too. What else would you do with an interpreter in the browser? What’s the point of having an entire IDE?
English
1
0
0
61
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen i would say if he claimed to plan to build the infra that allows agi (when its good enough that you do not need coding and asked people not to do coding) back in 2018 i would say this is the same vision to what hes doing now
English
0
0
0
24
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen the "build products" part is your addition
English
2
0
0
43
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen fyi its not a always a bad thing to change the vision .. it shows agility .. but its stupid to say he has not.
English
1
0
0
26
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen can you reference me the "we want to help people build products" in the period 2016-2022? the vision was teach people how to get into coding wo overhead .. this was replits offering for many years. the interpreter on the browser.
English
2
0
0
45
Adel Bibi
Adel Bibi@Adel_Bibi·
@Austen how is this the same single mission? he went from i want people to learn how to code without infra and local requirements to "people with coding experience are at a disadvantage"
English
1
0
0
77
Austen Allred
Austen Allred@Austen·
I’m not sure you could find many people in Silicon Valley who have spent longer going against a single mission
English
3
0
94
12.3K
Adel Bibi
Adel Bibi@Adel_Bibi·
@jamestanton Let x be uniform [0,1] and set A be the set of observing n heads. We want to compute P(A) = \int_x P(A|x)p(x) dx. But P(A|x) = P(n heads | x be prob of success) = x^n; and that p(x) = 1. Then, P(A) = \int_{0}^1 x^n dx = 1/n+1. For n =4, probability of getting 4 heads is 1/5.
English
0
0
3
690
James Tanton
James Tanton@jamestanton·
Much debate over the weekend about this classic gem: Your task: a) Select a number p between 0 and 1 at random. b) Manufacture four biased coins each with probability p of flipping HEADS. c) Flip those four coins. The Question: What are your chances of seeing four heads?
James Tanton tweet media
English
13
3
53
107.4K
Adel Bibi
Adel Bibi@Adel_Bibi·
@kenneth0stanley How is that different from taking current models and evaluating the generated proposed theory in the sciences? Oh, wait, we are already doing that.
English
0
0
1
87
Kenneth Stanley
Kenneth Stanley@kenneth0stanley·
It's interesting that we seek evidence of whether something is AGI in terms of whether a single model ("it") can do something like discover general relativity. But such discoveries are population-level feats, even if an Einstein is the final generator. The reason we could get one person in 1915 who could do that is that there were almost 2 billion diverse variations on the human mind at the time to choose from. To invest astronomical compute into a *single* model AI that could reliably do what Einstein did given knowledge up to 1911 is a vastly different problem than training 2 billion diverse models and finding the one (or it finding itself) with the right problem alignment. Humanity offers no precedent for cooking up a single model that can do anything the very best human can do in any field, though the subtlety of the distinction is easily overlooked.
Rohan Paul@rohanpaul_ai

Demis Hassabis’s “Einstein test” for defining AGI: Train a model on all human knowledge but cut it off at 1911, then see if it can independently discover general relativity (as Einstein did by 1915); if yes, it’s AGI.

English
51
30
236
33.5K
Adel Bibi retweetledi
Guohao Li 🐫
Guohao Li 🐫@guohao_li·
This is challenging to build safe multimodal OS agents. Check out @aichberger’s NeurIPS paper on how easy to attack multimodal OS agents with small malicious image patches
Lukas Aichberger@aichberger

Excited to present my paper '𝗠𝗜𝗣 𝗮𝗴𝗮𝗶𝗻𝘀𝘁 𝗔𝗴𝗲𝗻𝘁' today at #NeurIPS2025 🚀 We show how adversaries can craft and post small malicious image patches on social media to hijack multimodal OS agents ⚡️🤖⚡️ Come by today’s first poster session if you're around 👋

English
1
2
5
973
Adel Bibi
Adel Bibi@Adel_Bibi·
Fantastic work led by @aichberger . This work was also featured by the @sciam recently where we show a new type of model hijack that is as simple as posting an image on social media.
Lukas Aichberger@aichberger

Excited to present my paper '𝗠𝗜𝗣 𝗮𝗴𝗮𝗶𝗻𝘀𝘁 𝗔𝗴𝗲𝗻𝘁' today at #NeurIPS2025 🚀 We show how adversaries can craft and post small malicious image patches on social media to hijack multimodal OS agents ⚡️🤖⚡️ Come by today’s first poster session if you're around 👋

English
0
0
0
347
Adel Bibi
Adel Bibi@Adel_Bibi·
@yoavgo I'm confused. Still, what actual value does this bring beyond notation? Thinking of the reward as part of the environment allows us to standardized evaluation across agents to better evaluate learning algorithms as the objective is unified.
English
0
0
0
86
(((ل()(ل() 'yoav))))👾
(((ل()(ل() 'yoav))))👾@yoavgo·
I complain a lot about RL lately, and here we go again. The CS view of RL is wrong in how it thinks about rewards, already at the setup level. Briefly, the reward computation should be part of the agent, not part of the environment. More at length here: gist.github.com/yoavg/3eb3e722…
English
31
18
301
103.8K
Adel Bibi retweetledi
Peter Richtarik
Peter Richtarik@peter_richtarik·
yup
Adel Bibi@Adel_Bibi

It is just ridiculous to ask #ICLR ACs to go through a whole new batch of papers (12 in my case) to make a fair final judgment on the reviews without the final discussions and scores from the reviewers. At this stage, it feels like I am reviewing every single paper.

QST
0
1
9
3.2K
Adel Bibi
Adel Bibi@Adel_Bibi·
It is just ridiculous to ask #ICLR ACs to go through a whole new batch of papers (12 in my case) to make a fair final judgment on the reviews without the final discussions and scores from the reviewers. At this stage, it feels like I am reviewing every single paper.
English
0
0
29
6.7K
Adel Bibi
Adel Bibi@Adel_Bibi·
great opportunity and Jakob is an amazing collaborator; you will be getting a ton of support
Jakob Foerster@j_foerst

My Oxford lab (@FLAIR_Ox ) is hiring Phd students! If you are thinking of doing a Phd in blue-sky and -sort of crazy ambitious- ML and have a technically strong background and love to work with others, please consider all options for joining us: 1) Direct entry - deadline is the 1st of Dec AOE (ox.ac.uk/admissions/gra…) 2) AIMS CDT (ox.ac.uk/admissions/gra…) deadline on 27th of Jan 2026 AOE 3) EIT CDT (ox.ac.uk/admissions/gra…) deadline on the 7th of Jan 2026 AOE Student funding is a real constraint / concern in the UK (especially for overseas students) and by applying for these three programs you can maximize your chances of ending up in a very very special place.

English
0
0
3
692
Adel Bibi retweetledi
Fazl Barez
Fazl Barez@FazlBarez·
🚀New paper: "Chain-of-Thought Hijacking"! We found a universal jailbreak in Reasoning Models and worked with frontier labs to fix them! Our attack achieves >94% attack success rate against ALL leading proprietary models.🤯 1/7
Fazl Barez tweet media
English
2
27
88
16.5K

Keşfet

@Austen @jamestanton @kenneth0stanley @jamesaoldfield @aichberger @elonmusk @BarackObama @taylorswift13