David Schulz @TechySchulz

My ride from home to the University of Texas in Austin: Uber: $25.13 | 7 min wait | crappy car, unknown quality driver (usually bad) Lyft (promo): $19.23 | 4 min wait | same issues as Uber Robotaxi: $12.62 | 24 min wait | guaranteed new Model Y I’ll plan ahead and ride the…

keyboard_arrow_left Previous keyboard_arrow_right Next

The funny thing about the prediction that AI would be writing 90% of all code by now is that the prediction's failure distracts from the fact that AI adoption in code writing is actually extremely high, it was over 30% in December, 2024 according to one measure, with large impact

keyboard_arrow_left Previous keyboard_arrow_right Next

Anthropic co-founder, Jack Clark: AI isn't slowing down, and systems like those in "Machines of Loving Grace" look buildable by the end of 2026 that means a "country of geniuses in a Data Center" is likely by 2027, maybe even late 2026 progress hasn't plateaued; scaling laws…

What is the absolute best chat ui/ux or experience for LLMs? Anyone cracked this?

LLMs scaled rapidly because the internet gave us oceans of structured language data. But we lack equivalent datasets for the tacit knowledge hidden in scientific labs. A neat @Ben_Reinhardt proposal for "Unstructured Data Generation Labs" to change that: ifp.org/teaching-ai-ho…

There are probably other workflows and UXes that might work even better. Kind of a shame that the only place where AI interfaces are advancing rapidly are in coding (not surprising given who is making AI interfaces right now).

You meet a humanoid robot that passes every test of consciousness. It laughs, loves, dreams, and fears death. Do you treat it as a machine, or a person?

Batteries are underrated. Even in the short term, they’ll reshape industries, cities, daily life more than most people realize. They are about to do for energy what the microchip did for computing.

Dr Singularity @Dr_Singularity

5 days ago

Batteries are underrated. Even in the short term, they’ll reshape industries, cities, daily life more than most people realize. They are about to do for energy what the microchip did for computing.

13 58 349 20K 58

Download Image

A great article in @TheAtlantic about the extent of the penetration of the AI tools in education, high school and college in particular. The situation is pretty dire, but there is still hardly any serious initiative for a large scale reform that would be needed. All of our…

Surveillance tech is not making us safer – it’s putting innocent people at risk. This isn’t public safety. It’s automated injustice. Read more: eff.org/deeplinks/2024…

In a sane world we would understand that this is the mark of a civilization that has made a conscious choice against progress. Why did we stop going up the Kardashev scale?

An AI doctor emerges that never makes a mistake. No misdiagnosis, no malpractice, no death it could prevent. But patients start refusing it. They’d rather be treated by a flawed human who cares than a perfect machine that doesn’t. What does that say about us, that we value…

The new Codex is amazing, GPT-5 reasoning with practical unlimited usage, and you can select the reasoning level great work by @OpenAI At least in my testing it is every bit as good as Claud Code, even better in some tasks and even with the + plan you get almost unlimited usage

If a doctor ignores AI when it can raise diagnostic accuracy from 74% to 92%, that’s not “traditional medicine,” that’s negligence. We don’t excuse a surgeon who refuses anesthesia because “it wasn’t used before.” Same standard here. Not using AI in medicine is malpractice. Full…

Derya Unutmaz, MD @DeryaTR_

a week ago

If a doctor ignores AI when it can raise diagnostic accuracy from 74% to 92%, that’s not “traditional medicine,” that’s negligence. We don’t excuse a surgeon who refuses anesthesia because “it wasn’t used before.” Same standard here. Not using AI in medicine is malpractice. Full… https://t.co/wh81t8xFUF

20 68 486 84K 90

Today, my Model Y drove me 325 miles (5.5 hours) across three states, from New Hampshire to Upstate New York, on @Tesla’s FSD (Supervised) v13.2.9, with no interventions the entire way, even through rain. So, basically, I pressed 'Start,' and then a robot drove me for 5.5 hours.

We’ve been told “hard work makes you worthy.” AI blows that myth apart. A machine can now outwork the hardest worker alive… while you sleep. So if worth ≠ work… then what defines us? This is either the end of dignity or the rebirth of it.

ESET researchers discovered PromptLock, the first known AI-powered ransomware written in Golang with Windows and Linux variants that uses the gpt-oss-20b model from OpenAI via a proxy/tunnel to a server running the Ollama API to generate malicious Lua scripts on the fly from…

ESET Research @ESETresearch

a week ago

ESET researchers discovered PromptLock, the first known AI-powered ransomware written in Golang with Windows and Linux variants that uses the gpt-oss-20b model from OpenAI via a proxy/tunnel to a server running the Ollama API to generate malicious Lua scripts on the fly from…

56 675 3K 369K 1K

Download Image

Researchers at Singapore-based Sapient have developed a new AI architecture called the Hierarchical Reasoning Model (HRM) that mimics how the human brain processes complex tasks. Despite having only 27 million parameters, HRM outperformed leading LLMs like ChatGPT in the…

GPT-5 just outperformed licensed human experts on MedXpertQA, MedQA and USMLE benchmarks. Biggest leap is in multimodal reasoning where it beats doctors at combining text and images. Clinical decision support is about to change forever.

Google DeepMind just dropped Perch v2, an open-source AI model that analyzes nature sounds to identify endangered species from audio recordings -Discovered a new population of the elusive Plains Wanderer in Australia - 50x faster honeycreeper detection in Hawaii - Identifies…