AQGenesis

How DeepProbLog Could Revolutionize Autonomous Scientific Discovery What if AI could truly UNDERSTAND science, not just pattern-match? We explore a groundbreaking framework that teaches AI the rules of science, not just the data. Read more: aqgenesis.com/blog/deepprobl… A Thread ↓

Research domains are now available on Thinktica! 7 different research domains are now live: biomedical · computer science · physical sciences · legal · market intelligence · social sciences · generic

Big update in process! New sources are coming to Thinktica! arXiv, bioRxiv, pubMed, perplexity, Semantic Scholar and much more! The release will be made this week, we'll keep you updated 🚀

The first AI system that conducts scientific research autonomously 24/7 just found something remarkable. 4 genes independently evolved to extend lifespan in every long-lived mammal studied. Same pattern in 61 plant species. Full paper: zenodo.org/records/194656… The science 👇

Last week the world knew that @paul_conyngham sequenced his dog Rosie's genome and spent countless hours prompting ChatGPT, Gemini, Grok, and AlphaFold to design a personalized mRNA vaccine. The result? Her tumor shrank drastically, and she's back to running around. Even Sam Altman @sama called it an "amazing story" of how AI empowers individuals to act with the power of a research institute: x.com/sama/status/20… It's a powerful example of dedication. But what if you didn't have to spend weeks manually prompting different AIs, piecing together research, and verifying every claim yourself? One word: Thinktica thinktica.com @ThinkticaAI is an autonomous scientific research engine that does the heavy lifting for you. You ask one complex question, like finding a cross-disciplinary treatment for a specific genetic mutation and it runs for hours, building a deep, fully traceable knowledge graph. It decomposes your question, pursues every lead autonomously, and bridges fields like biology and materials science to find connections no single expert would see. Every finding is verified, every claim is traced to evidence, and there are no black boxes. Paul's story shows the raw potential of AI in medical research. Thinktica is the infrastructure that makes this level of deep, verifiable discovery accessible to EVERYONE! You set it going before bed. By morning, you have a fully traced research paper. With this step we're removing the ceiling on what one person can investigate. And best of all, Thinktica is powered by @eticaprotocol a decentralized blockchain network designed to fund and reward open-source medical research without intellectual property barriers creating a true revolutionary approach how medical research in the future will be made. Early access is open now. What will you discover? #AI #MedicalResearch #DeSci $ETI

Welcome @ThinkticaAI

Just tested thinktica.com with this exact question: “How liquid biopsy can be used to detect/characterize solid tumor development” I let the autonomous research agents run: 📹 First video 1min40s (20minutes research): Watch brand new research challenges and findings pop up live, highlighted in green with full sources. An AI Assistant was right there to guide, synthesize, or even generate content on demand. 📹 Second video 33s (one hour total research): The agents kept digging deeper and deeper autonomously. After just 1 hour of runtime, Thinktica autonomously wrote and formatted this complete scientific paper (full PDF link in following post): “Efficient Delivery and Precise Targeting of CRISPR Base Editors for Epigenetic Modulation in Human Primary Cells” Thinktica Research – March 21, 2026 Interestingly, the agents smartly pivoted into CRISPR base editors for epigenetic editing, which is actually super relevant: modern liquid biopsy relies heavily on detecting aberrant DNA methylation patterns in cfDNA to identify and characterize solid tumors. The system went straight for next-gen tools in the field 🤯 These two videos capture only 60 minutes of autonomous research. Just imagine what this system could uncover in days or weeks of continuous deep research… 🤖🧬 This is the game changer we’ve been waiting for. Autonomous agents that don’t just search, they investigate, connect dots, and publish real papers while you sleep! Powered by @eticaprotocol $ETI #DeSci

@adibvafa Ok thank you for info. Do you have an api available and if not do you plan on making one? And do you have any data about hallucination rate of your system, how reliable is it at the moment? Thanks

@AQGenesis yes! you can find all the assets here bioreason.net/code

Proteins can now talk. Introducing BioReason-Pro, the first reasoning model for protein function. A thread🧵

@adibvafa Nice work! Is it open source?

You can talk to it here! bioreason.net Paper: bioreason.net/paper Code: bioreason.net/code Data: bioreason.net/data Weights: bioreason.net/sft , bioreason.net/rl Catalogue of 240,000+ predictions: bioreason.net

The future of open medical research just went autonomous. @eticaprotocol + thinktica.com = the first full-stack DeSci engine that doesn’t just fund science… it does the science! No patents barriers or Intellectual Property restrictions! Next few weeks will be insane!

Scientists just tested a vaccine that could fight any cancer. A team at the University of Florida has developed an experimental mRNA cancer vaccine that could teach the immune system to fight many types of tumors — not just one. Instead of zeroing in on a single cancer marker, this vaccine trains the body to recognize and attack cancer broadly, much like it responds to a virus. In lab tests, when paired with existing immunotherapy drugs, the vaccine triggered a strong immune reaction that shrank stubborn tumors in mice — including cancers that had resisted treatment before.

The Future Ecosystem: From Capital Accumulation to Scientific Breakthroughs • BTC — reserve / gold • ETH — compute / world computer • XMR — anonymity / privacy gold standard • ETI — open science / medical progress $ETI #Blockchain #DeSci #Web3

Building AI infrastructure in medicine is important. Introducing CancerLLM: a domain-specific large language model trained entirely for oncology. Built on 2.7 million clinical notes + 515,000 pathology reports across 17 cancer types, this 7B-parameter model outperformed existing LLMs by ~9% F1, hitting: • 91.8% for cancer phenotyping • 86.8% for diagnosis generation And it does it with far lower compute than massive general models. This is an example of oncology-native language infrastructure. nature.com/articles/s4174…

AI Just Did What Physicists Thought Was Impossible. GPT-5.2 helped derive a brand new result about one of the fundamental forces of the universe. Not automation. Not productivity gains. Actual scientific discovery. We are officially living in the age of machine-driven physics. Most people still have no idea what just started.

Open source will become a Must have for 21th century AI economy

@DeryaTR_ @steipete @openclaw Hi Derya, also check Thinktica it’s a new open source framework specialised in 24h/7 AI autonomous research agents. With the framework you can get an autonomous researcher running in just few command lines Doc: thinktica.gitbook.io/thinktica-cli

ClawdBot is the new rage in agentic AI! This should be the next version of Claude Cowork. I’ll test it on an idle Mac next week. Be cautious about installing it on your main computer; that can be risky as it's not security-grade yet. Kudos to @steipete for developing @openclaw!👏

Germany’s Breakthrough in Regenerative Medicine: Self-Regenerating Synthetic Cartilage In a groundbreaking achievement, a team of regenerative medicine researchers in Munich has developed synthetic cartilage that doesn’t just replace damaged joints, it actually regenerates itself inside the body. This innovative technology could spell the end of traditional joint replacements for conditions like arthritis, injuries, and worn-out knees. The synthetic cartilage, made from a bioresorbable hydrogel with stem-cell attractors, functions like scaffolding when implanted into the body. It encourages nearby stem cells to migrate, attach, and grow into real cartilage, restoring the cushioning and flexibility of joints without the need for full surgery. In animal trials, the damaged knees were restored to 90% of their original function within just 4 months. Remarkably, the synthetic material gradually fades away, leaving behind fully grown, natural cartilage. Unlike current implants, this material doesn’t wear out over time, and there’s no need for future joint replacements. The procedure, which can be done through keyhole surgery in under 45 minutes, offers quick recovery times weeks instead of months and eliminates the need for extensive physical therapy. The German health ministry has fast-tracked clinical trials for this revolutionary technology, and if approved, it could potentially eliminate the need for artificial knee replacements for good. #RegenerativeMedicine #SyntheticCartilage #JointRegeneration #HealthInnovation #StemCellTherapy

It's not only incredibly impressive that GPT-5.2 Pro has now achieved 31% of the FrontierMath benchmark (no overfitting), but I find it even more impressive how it directly solves new problems and how impressed the brightest mathematicians are with GPT-5.2's capabilities. I'm incredibly excited for GPT-5.3, should it be released soon. Until then, be sure to read the professors' reports.

Aging is a solvable problem that should get fixed with Open Source research. If only billionaires get aging cures then the world will start to look like Elysium @eticaprotocol

@CryptooRay 100% few people understand it yet, Etica holders are the new Bitcoin 2011 era like holders. You’re extremely early to understand this change, time will reward your foresight in this ocean of shitcoins you managed to identify the real future game changer

@AQGenesis X provides real-time collective intelligence → Thinktica provides intelligent infrastructure → Etica provides continuous incentives. Perhaps elon musk will consider this closed loop in the future.

Thinktica is coming. The first open-source framework for building autonomous AI research agents. What takes 20000+ lines of code today will take 150. Documentation is live. Launch is imminent. 🧵 Thread below.