Mamad Ahangari
116 posts
@centrum_blue
PhD in statistical genetics. Scientist at @herasight. Interested in decentralized AI (τ) and late antiquity history. Opinions my own.
Publishing the outcome of several months of research on egg freezing and IVF, some of which shocked me. Most American women aren't told there are different versions of freezing/IVF protocols— and the version they're offered by default is often the most aggressive. In much of Europe and Japan, women get half (or less) of the same medication for the 'same' procedure and outcomes. The assumption baked into American fertility medicine is that more is better. More eggs help up to a point; past it, you're buying more hormone exposure and more risk — not necessarily more babies. Long-term safety data on healthy women taking these protocols repeatedly at these doses doesn't yet exist. The studies that look reassuring follow women for 5-10 years, not 30. Full essay linked below. I tried to write the overview I couldn't find while making these decisions.
Klausner et al. makes a valuable contribution to modeling polygenic embryo screening (PES) efficacy in real-world IVF data. However, there are some important methodological and conceptual limitations worth discussing 🧵
$TAO Making a difference! ⬇️⬇️ MINOS SN 107 👀👀 Alagille syndrome is a rare genetic disorder affecting roughly 1 in 30,000–50,000 live births. It causes problems in multiple organs, especially the liver (due to missing or reduced bile ducts leading to cholestasis, severe itching, and growth issues), heart, skeleton, eyes, and blood vessels. Prognosis varies widely — overall survival is around 80–90% into adulthood, but 20–30% of patients eventually need a liver transplant, and native liver survival drops to about 40% by age 18. Current treatments focus on symptom management (e.g., medications like maralixibat or odevixibat for itching, nutritional support, and surgery), with liver transplantation as the main option for severe cases. There is no cure that fixes the underlying JAG1 mutation. Why better analysis in this region helps: Accurate, low-cost detection of JAG1 variants enables earlier diagnosis, personalized monitoring, and timely interventions — potentially reducing complications, improving quality of life, and lowering the need for costly transplants or long-term care. This is exactly where Minos’ decentralized AI validation adds real clinical value. Minos’ value mining near the JAG1 (chr20:11M-16M) region: Imagine trying to read a critical page in a book, but some words are smudged or in hard-to-read font. Traditional DNA analysis tools often struggle in this specific JAG1 region — it’s technically “noisy” and error-prone. What Minos is doing: • They’re crowdsourcing thousands of decentralized AI “miners” to double-check and perfect the reading of this exact region. • Goal: Reach near-perfect accuracy (F1 score 0.95–0.99), matching or beating top tools like DeepVariant and DRAGEN. Why this matters (the real value): • Mutations here cause Alagille syndrome — a serious condition affecting the liver, heart, bones, and more. • Better accuracy = faster, more reliable diagnoses → earlier treatment, better patient outcomes, and fewer unnecessary tests. • For researchers and clinics: Cheaper re-analysis of existing DNA data (saving $10–$100+ per sample in this region) instead of expensive re-sequencing. • Long-term: This proves decentralized AI can deliver clinical-grade genomics at lower cost, opening the door for broader, more affordable precision medicine. In short: Minos is turning a hard, high-stakes genomic “problem zone” into a reliable, low-cost strength — directly helping rare disease patients while showing how decentralized AI can compete with (or beat) big centralized tools. This small region is a perfect showcase for the whole project’s potential. #Genomics #PrecisionMedicine #AlagilleSyndrome #JAG1 #Bittensor #DeAI #BioTech #HealthTech #RareDisease #DNA $TAO
Looks like I might be able to vibe-mine Minos SN107. They have the key mining component -- MinosVM -- set up as a rentable on Targon SN4.
Bill Gates: "6% of global emissions are cows… You can either fix the cows to stop them doing that, or you can make beef without the cow."
Lium will become the first agent-first compute provider. We are leaning in heavily to the agent revolution. Here's how we make changes to lium now: We prompt our agents: "using this crypto account, create an account on lium, fund it, get an api key, rent some GPUs, and serve Kimi K2.6 with sglang (or another gpu-required task). When done, create a report for any issues you encountered with using lium and how it could have been faster or easier" Then we constantly improve the agent experience so that it's seamless to tell your agent to do any task, and it will just spin up lium gpus in seconds to solve it effortlessly and without any human intervention needed. We have recognized that SAAS companies need to pivot to agent-first software in order to keep up, and lium is the one leading the charge on that for cloud compute.
Introducing GPT-5.5 A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done. Now available in ChatGPT and Codex.
🚨 Over 1 billion rows of psychiatric genetics data. Now on Hugging Face. ADHD. Depression. Schizophrenia. Bipolar. PTSD. OCD. Autism. Anxiety. Tourette. Eating disorders. 12 disorder groups. 52 publications. Every GWAS summary statistic from the Psychiatric Genomics Consortium. Before: wget, gunzip, 20 minutes debugging separators, repeat 50 times. Now: one line of Python.
Today the worlds most powerful genetic predictor of IQ, CogPGT, has been published in the peer reviewed journal Intelligence and Cognitive Abilities. When used for embryo screening, it can substantially boost expected IQ of future offspring. Read on for the scientific details!
