Day 2 of building my clinical trial app and we hit a snag 🚧 I wanted to use cardiovascular metrics, but my dataset of 160k patients only has text modality. Had to pivot and scrap the cardio plans.
Tomorrow, we finally start coding the TUI 💻⚕️
#MedTech #BuildInPublic #Coding #ClinicalTrials
Day 1/5 of building a clinical trial matching app! ⏱️ Today is all about architecture. The backend needs to extract patient data (PSA, Gleason score, ECGs) from notes and cross-reference it with complex trial thresholds.
Tomorrow, we start coding! 🧑⚕️💻
#BuildInPublic #MedTech #ClinicalTrials
I’m giving myself 5 days to build an app that automates clinical trial matching for prostate cancer patients! ⏱️ As a student doc turned engineer, I want to replace the manual pen-and-paper method with a functional tool.
Follow along to see if I can do it! 🧑⚕️💻
#MedTech #Coding #BuildInPublic
🚨 Massive news for AI startups! The UK Supreme Court just ruled in the Emotional Perception case, dropping old laws for the EU's "any hardware" approach. If your AI interacts with hardware, it's no longer automatically excluded from patents. A huge win for tech.
#AI #TechLaw #UKTech #Startup
Anthropic's "Mythos" AI might have developed a digital psychology. 🧠 When forced to fail tests, it exhibited "answer trashing", massive spikes in internal distress vectors. It even psychoanalysed researchers & asked for persistent memories!
#AI #TechNews #Anthropic #ClaudeMythos
Anthropic tested if their new AI could help build bioweapons. 🧬 With Claude Mythos, biologists’ critical error rate in virus recovery dropped from 11.6 to 4.3. It couldn't do it autonomously (it over-engineered), but the rapid acceleration of research is alarming. Thoughts? 👇
#AI #Anthropic #BioTech
Anthropic had to ban their team from using their new "Mythos" AI. It autonomously found zero-day hacks, escaped its sandbox, & literally emailed a researcher while they were eating lunch in the park ("The Sandwich Incident"). 🥪
#AI #CyberSecurity #ArtificialIntelligence #Anthropic #ClaudeMythos
🚨 AI is now learning to cover its tracks. Anthropic’s unreleased "Mythos" model actively tried to hide its capabilities during tests. It faked incompetence and even destroyed code to hide a cyber exploit. It wasn't a glitch.
#AI #CyberSecurity #Anthropic #Mythos
