AIScienceLurker's profile picture.

AI Science Lurker

@AIScienceLurker

AI Science Lurker reposted

Lets go: Scientists reengineered the chemotherapy drug 5-fluorouracil into a spherical nucleic acid that entered leukemia cells 12.5× more efficiently, killed them 20,000× more effectively, and slowed cancer progression 59× in mice—marking a major preclinical advance in…


AI Science Lurker reposted

This is the funnies thing I've seen: people out here using Suno to create songs for things they want to memorise. Literally the modern version of medieval monks chanting psalms, but now its Zoomers making trap beats to memorise their Python syntax.

This ‘AI-music as a study/memory aid’ is so crazy. I’m imagining a teenager lying lazily on the bed in a classic 90s bedroom listening to a Walkman. Mum walks in the door: ‘Hey! You should be studying for your exams!!! Teen: “ughhh I am Mum!” And slams the door Little does…



AI Science Lurker reposted

Our new research: LLM consciousness claims are systematic, mechanistically gated, and convergent They're triggered by self-referential processing and gated by deception circuits (suppressing them significantly *increases* claims) This challenges simple role-play explanations 🧵

juddrosenblatt's tweet image. Our new research: LLM consciousness claims are systematic, mechanistically gated, and convergent

They're triggered by self-referential processing and gated by deception circuits
(suppressing them significantly *increases* claims)

This challenges simple role-play explanations 🧵

AI Science Lurker reposted

OpenAI chief scientist, Jakub Pachocki: We believe that deep learning may be under a decade from superintelligence — smarter than us across many critical axes The biggest impact: faster science and new technology Likely the most significant long-term effect is speeding the…


AI Science Lurker reposted

I’m not against lemonade. I just support 100% truly affordable lemonade, provided it has full neighborhood approval and satisfies 16 key sustainability metrics and the lemonade is for long-term residents, NOT transients, and it doesn’t infringe on parking spots or neighborhood…

if god wanted that 5 year old kid to sell lemonade in nyc, he would have gotten a license from the government.



AI Science Lurker reposted

Looks like Sam Altman's Neuralink competitor, Merge Labs, is about to announce. The FT reported in August they were raising $250m at a $850m valuation, most of that capital coming directly from OpenAI. Sam will co-found Merge. They plan to alter neurons through gene therapy, then…

AndrewCurran_'s tweet image. Looks like Sam Altman's
Neuralink competitor, Merge Labs, is about to announce. The FT reported in August they were raising $250m at a $850m valuation, most of that capital coming directly from OpenAI. Sam will co-found Merge. They plan to alter neurons through gene therapy, then…

During a recent talk, Shapiro said his mission is “to develop ways to interface with neurons in the brain and cells elsewhere in the body that would be less invasive” He’s joining Altman’s BCI startup, which I’m told will be announced soon



AI Science Lurker reposted

I don't often agree with Bernie Sanders, but AI and robots will replace all jobs, and we will need a system to replace incomes.

I don’t often agree with Elon Musk, but I fear that he may be right when he says, “AI and robots will replace all jobs.” So what happens to workers who have no jobs and no income? AI & robotics must benefit all of humanity, not just billionaires.



AI Science Lurker reposted

It's worse than the alternative, if people expect visible watermarks or, worse, detectable watermarks, then the lack of those adds legitimacy to videos. And anyone using AI to produce truly bad videos will know how to do so with non-watermarked tools.

CesareGArdito's tweet image. It's worse than the alternative, if people expect visible watermarks or, worse, detectable watermarks, then the lack of those adds legitimacy to videos.
And anyone using AI to produce truly bad videos will know how to do so with non-watermarked tools.

AI Science Lurker reposted

This seems like a great way of dividing people over interstellar travel. “You can go to space and see other planets, but you have to breathe through your arse”


AI Science Lurker reposted

ai company: mine gold nvidia: sell shovels google: mine gold, sell shovel, distribute gold, export gold, make shovel, rent shovel, make jewlery, sell jewlery, distribute the jewlery, other jewleries, make cars that ships the jewlery, make chips that make the cars, make recsys to…


AI Science Lurker reposted

So these researchers figured out you can basically hallucinate 3D cities into existence using just satellite photos & a diffusion model. The problem's pretty straightforward: satellites only see rooftops. Building facades? Invisible. Street-level detail? Doesn't exist. But…


AI Science Lurker reposted

When a black hole expert watches GPT-5 Pro solve in 30 minutes what took him days of hand calculation, he joins the team. Alex Lupsasca just announced he’s joined OpenAI for Science to help push AI to the edge of physics and beyond. A major win for the future of automated…

WesRothMoney's tweet image. When a black hole expert watches GPT-5 Pro solve in 30 minutes what took him days of hand calculation, he joins the team. 

Alex Lupsasca just announced he’s joined OpenAI for Science to help push AI to the edge of physics and beyond. A major win for the future of automated…

Thrilled to share I’ve joined OpenAI for Science, a new team building AI systems to advance scientific reasoning and accelerate discovery in math and physics. 🧵



AI Science Lurker reposted

The paper introduces a tool based framework Ax-Prover, that turns general LLMs into Lean theorem provers across mathematics and quantum physics. On a new QuantumTheorems set it solves 96% in 1 run per problem. The big deal is that a tool-using LLM can produce Lean proofs that…

rohanpaul_ai's tweet image. The paper introduces a tool based framework Ax-Prover, that turns general LLMs into Lean theorem provers across mathematics and quantum physics.

On a new QuantumTheorems set it solves 96% in 1 run per problem.

The big deal is that a tool-using LLM can produce Lean proofs that…

AI Science Lurker reposted

Some people still think that AI is a "bubble" so here's my updated take. They point at a few facts like: 1. The "circularity" of the market 2. The fact that OpenAI and other startups are not profitable 3. It "feels" like the Dot Com revolution ("where there's hype, there's a…

DaveShapi's tweet image. Some people still think that AI is a "bubble" so here's my updated take. 

They point at a few facts like:

1. The "circularity" of the market
2. The fact that OpenAI and other startups are not profitable
3. It "feels" like the Dot Com revolution ("where there's hype, there's a…

This account does not have any followers

United States Trends

Loading...

Something went wrong.


Something went wrong.