Dr. ELL retweeted
Hello everyone, I’m thinking about doing a research internship at a company or startup next year. Ideally, I’d be looking for something later in the year, around August, although the timing is flexible. I’m open to opportunities in the UK, US, EU, or Canada. I’m emphasising research because, although other types of internships, such as software engineering, are interesting to me, they’re not the best fit for my background.
My interests are relatively broad, mostly around AI for science, mathematics, and code, as well as understanding large models.
My current research is centred on topological and geometric machine learning. Much of my work focuses on developing probabilistic models over structured geometric domains such as graphs, simplicial complexes, and manifolds. It’s a happy mix of probabilistic modelling, differential geometry, and numerical analysis. I’m particularly interested in vector field applications and how to discretise domains so that probabilistic models preserve their underlying geometric structure.
Some of the questions I’d be interested in exploring include how to develop machine learning methods that respect the geometric and topological structure of a problem, how to understand the behaviour of large models (and apply them to scientific problems), and how to use AI to develop new algorithms and advance mathematics.
I’m not necessarily set on doing another internship (I’ve already done one), but I’d love to explore opportunities if there’s a good match between my research interests and what a team is working on. If anyone knows of teams, companies, startups, or researchers who might be a good fit, I’d be very interested in hearing about them.
I don't think it's that crazy even if you have a firm stance that AI models isn't and never will be conscious. Even taking the extreme stance that they are stochastic parrots, the fact they are trained on user interactions, it just makes sense that they want to reduce learned cruelty behaviour that they will have to later clean up.
On the Anthropic side, I do think many people there believe AIs could be conscious, and almost certainly believe that even if it isn't conscious now, it could be in the future (with greater complexity, with new hardware), and having this in its logs, training data, history etc is probably not a good start for developing a conscious being.
Dr. ELL retweeted
"How do single-celled algae swim toward light sources" is the kind of project that gets DOGE'd—how does this help Americans?
20 years later, the underlying protein is a central tool in human neuroscience.
We're a rich country, we should be funding way more basic science.
Dr. ELL retweeted
Every academic knows that citations aren’t everything. We know this because we check them regularly.
Logged into "asiandating . com". my guy....
NEW: We’ve spent months reporting on a Chinese intelligence effort to penetrate the Federal Reserve. It’s a fascinating and important saga, revealing the lengths that Chinese spies are willing to go to get any scrap of information about the Fed. It features lies, betrayal, nude photos, extortion, false identities and document leaks. And it doesn’t end at all the way you might expect: cnbc.com/2026/09/30/john-rog…
Holy slop video for a launch.
If a picture is worth a thousand words, your camera roll is worth millions.
That’s why we built Dazzle, an AI assistant that turns your camera button into a "get it done" button. No blank prompts, just instant context and action. 🚀
Read the full launch post: dazzle.ai/blog/ 🎉
The fact that both husband and wife are in on it, and not just one of them. Real commitment to the act there.
Trans NYT exec threatened to kill wife, scared their kids with chilling book before his in-laws shot him dead: docs trib.al/3BBWgnq
Every time I see this and I think it’s a valid answer. You can cook rice in a pot, a rice cooker, a pan (paella/risotto), a clay pot etc. how one cooks rice has a huge effect on how it tastes
Anyone know if de-anonymization refusal is an important safeguard being worked on by OpenAI? It refuses attempts to use it de-anonymize someone but there's some pretty gaping workarounds and failures.
Never understood the "lobster too buttery" saying. Steak too juicy makes sense, it's a combination of meat quality and chef skill to have it juicy. Butter is just butter, you can add more if you want to and relatively cheaply.
The biggest lobster tails in Vegas are at The Golden Steer at $110 each. What do you think?
📍308 W Sahara





























