Anaconda Introduces Next-Gen CLI with Governed Packages

Today, Platformatic released "Project Destino".   The project started with Paolo Insogna joking about it... but obviously, Matteo Collina and Luca M. took him way too seriously.   𝘚𝘱𝘰𝘪𝘭𝘦𝘳: 𝘗𝘢𝘰𝘭𝘰 𝘪𝘴 𝘯𝘰 𝘭𝘰𝘯𝘨𝘦𝘳 𝘫𝘰𝘬𝘪𝘯𝘨 𝘸𝘪𝘵𝘩 𝘓𝘶𝘤𝘢 𝘢𝘯𝘥 𝘔𝘢𝘵𝘵𝘦𝘰 𝘢𝘧𝘵𝘦𝘳 𝘵𝘩𝘪𝘴 𝘱𝘳𝘰𝘫𝘦𝘤𝘵!   But anyway...   Behind this funny project... there is an extremely important value for all the enterprises using Node.js out there, and this is exactly what Luca and Matteo saw in the joke.   If Node.js can run Doom at 35 fps with sound, it can call the legacy C library that enterprises are afraid to touch. Here is a quick demo on how we made DOOM run in the terminal. Full article in the comments! ⬇️ https://lnkd.in/gyXX9Xc9

🚨 Important notice🚨 LAVA is a beta release and not all artifacts in the LEAPPs are currently supported. If you expect to see something in LAVA and it is not there check the LEAPPs' html report on it. The plan is to eventually have all artifacts be LAVA supported. Anyone that wants to help update current artifacts for LAVA compliance is welcome to assist us. Here is a video explaining how to update the artifacts for LAVA support: https://lnkd.in/eVrtXTJd #DigitalForensics #MobileForensics #DFIR

Anthropic : Well, You think Claude can't build and update live backend systems from prompts? Thomas Hansen: Ainiro builds and updates live backend systems from prompts, something Claude does not do. Thomas Hansen: Claude focuses on text generation and assistant workflows. It does not ship with a backend application engine like Ainiro. Anthropic : I see 😏

Just a year ago IOTA officially unveiled Rebased - a complete overhaul of the IOTA Layer 1. With that upgrade, IOTA has become one of the oldest Layer 1's without any technical debt. Over the past year the team has worked on many improvements to protocol. Today, with Starfish, IOTA has completed the biggest upgrade and improvement to the IOTA mainnet. The improvements are clearly visible and the Starfish consensus will make IOTA more resilient, scalable and performant.

Wrapping up strong in London! Today marks the final day of the OGC Builder Days Code Sprint, bringing the community together for another day of practical work sessions, technical discussions, and demos.  Day 3 features sessions and discussions on: • OGC API - Processes with pygeoapi and Prefect  • A Case Study in Vibe-coding OGC CSAPI  • A Tutorial and Walkthrough Implementing Binary Encodings with OGC API - Connected Systems  • WebVMT: GIMI Video Metadata on the Web  • A Tutorial and Walkthrough Implementing Pub/Sub MQTT with OGC API - Connected Systems  • A Tutorial on Automated Tipping and Cuing Between A Satellite and Drone Using OGC API - Connected Systems A huge thank you to everyone participating, presenting, and contributing throughout the sprint at Geovation in London. Want to keep the conversations going? Join us next at OGC Connect Helsinki from June 1–4, 2026 for the 135th OGC Technical Meeting. Register now: https://lnkd.in/gcKM8a9A #OGC #CodeSprint #Geospatial #Interoperability #OpenStandards #OGCAPI #GIS #DeveloperCommunity 

Boardroom-grade massing data for the Belgrade Skyline development. We embed the hero asset's native 2.5D geometry directly into the grid, mapping precise proximity to key state infrastructure and transport nodes for institutional capital evaluations.

In the second post of our Time Travelogue series, Mark Erikson uses Replay MCP to analyze and fix a set of bugs that Nadia Makarevich posted about recently. True to form, Mark takes us on a journey into the gory details of his rigorous, experiment-driven approach, and in the end he not only solves the bugs, but also uncovers ways we can make Replay even better at what it does. Whether you are debugging locally, or want to automatically get root-cause analysis and suggested fixes posted as comments on your PRs as part of your CI/CD pipeline, we’ve got you covered. Link to post in the comments

Our AI-native agent, the Guardener, is an engineer's new BFF 🤗 ✔️ Intelligent context-aware conversion ✔️Builds and tests layer by layer ✔️Deeper context, deeper telemetry via multiple deployment options And now you can see it in action. Check out Mandy Hubbard migrating a real Dockerfile from upstream to Chainguard Containers at the link in comments 👇  

Qwen3.6 35B was strong enough that I had to make the benchmark harder. The original suite was getting too clean. Too many passes. Not enough room for agents to get exposed. So I made new tasks with longer horizons, heavier repo navigation, trickier terminal workflows, and more ways to quietly fail after looking competent for most of the run. Villani still won. Headline result: Villani solved 55/57 Claude Code solved 50/57 That’s 96.5% vs 87.7% on the exact same task set, using the exact same backend model. The gap got wider where it actually matters. On hard tasks, Villani got 24/25. Claude got 20/25. On terminal tasks, Villani got 11/11. Claude got 8/11. On the hard and extreme suites, Villani went 6/6 and 6/6. Claude went 4/6 and 5/6. There were 5 Villani-only wins. Claude had 0 tasks it solved that Villani missed. Same model. Same tasks. Different runner. That is the story. A lot of people still act like the model is the whole product. It isn’t. Once the base model is strong enough, the bottleneck moves into planning, localisation, verification, edit discipline, recovery, and terminal behaviour. That is where the real gap shows up. The model is the engine. The runner decides whether the car stays on the road. GitHub: https://lnkd.in/gxUUHAvB

Anthropic is not acquiring Hey API, an SDK generator that has powered no Anthropic SDK since the earliest days of their API.