The next leap in AI won't come from more internet text. Models are hitting a ceiling on recycled data. What moves the needle is Frontier Data — the kind that doesn't exist anywhere yet and can't be scraped: → expert knowledge from domains that never made it online → edge cases that simulators can only approximate → robot manipulation footage from real-world environments → verified onchain address labels That's what Codatta contributors are building, task by task.

Teaching a robot to pick up objects is hard. Teaching it to operate them is harder. Our Appliance-Knobs Dataset on @huggingface focuses specifically on this type of fine-grained interaction. It features detailed visual data tailored for capturing the subtle geometric and functional variations of rotary controls. What makes it different: 1️⃣ Multi-Angle Views: Paired images (front & side) for every knob, giving models the multi-perspective data needed for robust 3D shape estimation. 2️⃣ Specialized Focus: A deep dive into electrical appliance knobs—an underrepresented class crucial for fine-grained object understanding. 3️⃣ Precision Ready: Optimized for state recognition and exact knob angle/position estimation. Built for: Multi-View Object Recognition | 3D Shape Reconstruction | Generative AI Training. 🔗 Download the dataset:

We’re launching the Latch MCP and announcing its availability within Claude Science, Anthropic’s new AI workbench for scientists. AI for biology requires agent-native infrastructure: systems where agents can store, process, and visualize large molecular datasets from the interfaces scientists already use. Biological analysis workflows often require substantial compute. Retries or incorrect long-running tool calls can quickly inflate workflow time and cost, especially when analyses take hours or days to complete. These tools should also be curated with appropriate parameters and agent-readable documentation, ideally provided by the original assay developer, to support correct use across many complex scientific contexts. At Latch, we’ve seen customers of our Solution Provider partners, including TakaraBio, Vizgen, and AtlasXOmics, use both the Latch Agent and external harnesses like Claude Code and Cursor to accelerate analysis of their data. The Latch MCP is a remote MCP server that securely connects agent harnesses to the Latch platform, giving agents access to verified bioinformatics tools built and maintained by kit and instrument providers. Agents can navigate data on Latch and launch existing deployments of validated bioinformatics workflows to analyze that data.

With X's hosted MCP server available, you can now connect any MCP-compatible AI tool, including Grok Build, Cursor, Claude, VS Code, and others to run MCP servers on Asmora. Through the open-source xurl bridge, OAuth is handled automatically and fresh Bearer tokens are injected on every request, allowing you to securely search the full X archive, look up users, manage bookmarks, fetch trends and news, draft Articles, and access X developer documentation all using your own X account permissions and without the complexity of manual MCP setup. Connect to and run hosted MCP servers through Asmora via our terminal today:

Announcing the hosted X MCP. Agents now have access to the best real-time information source in the world. Connect Grok, Cursor, or any MCP-compatible AI tool to the X API without any setup! Check it out here:

BioMCP by @imaurer now enables biomedical agents to rapidly explore the treatment landscape for KRAS G12C-positive NSCLC using just seven CLI commands. By integrating data from 10+ biomedical sources, it retrieves therapeutic evidence, maps approved and investigational drugs, reviews FDA safety data, analyzes real-world adverse events, identifies recruiting clinical trials, and summarizes the latest resistance research all within a reproducible, end-to-end command-line workflow. Explore BioMCP on Asmora today:

Biomedical agents can understand the treatment landscape for KRAS G12C-positive NSCLC by chaining together 7 BIOMCP commands. The below terminal-based screencast demonstrates the BioMCP commands that: * Orients across 1,741 known variants and 84 clinical trials * Pulls CIViC and ClinVar therapeutic evidence for G12C * Maps 5 drugs targeting KRAS (sotorasib, adagrasib, and 3 investigational) * Reviews sotorasib's FDA approval history and safety profile * Checks 2,465 real-world adverse event reports from FDA FAERS * Finds 62 recruiting trials for KRAS G12C NSCLC * Retrieves the latest research on resistance mechanisms 7 entity types. 10+ data sources. Under two minutes. Drugs, evidence, safety, trials, resistance -- the complete picture from a single CLI. Full walkthrough with every command and output: Documentation: GitHub Code repo: #BioMCP# #PrecisionOncology# #KRASG12C# #NSCLC# #AIAgents# #ClinicalGenomics#

Excited to share new work with @NVIDIAAI: we benchmarked 10 BioNeMo NIM skills across three @AnthropicAI Claude models, ~830 controlled runs. The result: skills don't make the models smarter, they make delivery reliable. On hard calls, a model ~5x cheaper became more reliable than the frontier baseline.