I built a new version of flat.social, a spatial meeting app where you can hang out with your remote colleagues and schoolmates. It's 3D now so y’all can jump around with pet robots and fly on balloons and helicopters! Try it at flat.social

Today we are talking with Vignesh Baskaran, the CTO and co-founder of @hexoai, about teaching AI agents to improve themselves. Vignesh has been training neural networks since 2012, back when he was still called a data scientist. Then he became an ML engineer and now an AI engineer, though he says the underlying work has never really changed. It's to figure out how to make a system behave the way you intend it to. He built the litigation search engine that Google itself became a customer of, and now he's chasing something new, agents that rewrite and retrain other agents without a human in the loop. We dig into Sia, the meta-agent at the center of Hexo's research, and why improving an agent means touching both its harness and its actual model weights, not just one or the other. We talk about proxy evals for when you don't have much to ground truth. The Darwin-Gödel machine and why formal verification is too strict a bar for anything commercial. How Hexo's work echoes DeepMind's Alpha lineage from AlphaGo to AlphaEvolve, and the spectrum from clearly verifiable to totally subjective tasks? Why VAE evals are quietly wrecking agent quality across the industry, and a great story about an agent that discovered a customer's own eval file was silently corrupted, something buried in hundreds of thousands of traces that no human would have caught. Watch on YouTube: piped.video/SEYmUH_Ae3U @tweetvbaskaran

DARPA Lift Challenge — T–15 Days darpa.mil/research/challenge…

We are hiring part time podcast hosts for Software Huddle. Please reach out to team@softwarehuddle.com if you are interested.

Most people think the AI transition is happening inside AI companies. It's not. The whole ecosystem is accelerating. @grinich joined @KentBeck on Still Burning to talk about what enterprise infrastructure reveals about where tech is headed. piped.video/Kh24KYFfH5Q?si=HWJX…

Glauber from @tursodatabase talks about his AI stack✨ @glcst

➡️@valkey_io is the new Redis.

nobody talks about token economics seriously and Gemini 3.5 Flash is exactly why they should. quick thread on what's actually going on with AI costs and why it matters more than benchmarks

Catherine Jue talks about 3 things her team at @usekernel looks for in new hires. Visit kernel.sh to check out how Kernel is building Open Source infra for AI Agents.

Hour 24 - won VillageHacks at ASU Day 10 - got into Momentum (@Devlabs_club) Day 20 - first demo vid on X here's the problem we're solving- AI agents have no memory between sessions. but it's worse than that. even mid-session, context gets compacted. an infra decision made 3 months ago, a dead end you already hit, quietly erased as the chat grows, so when that same infra gets a bug today, your agent is reasoning blind. no memory of why it was built that way. no memory that the obvious fix was already tried. starting from zero. again and again. trace is the intent-aware memory layer your agent was never built with. #buildinpublic #AI #devtools #aimemory #Devlabs #startup

AI and Proactive Reliability with Kolton Andrus Today we're talking with Kolton Andrus, the Founder and CEO of @GremlinInc, about what happens to reliability when AI is writing most of the code. Kolton helped build the Chaos Engineering practice of both Amazon and Netflix before starting Gremlin. In our conversation we talk about scar tissue, the intuition engineers develop from being woken up at 3:00 AM to fix production outages, and how AI doesn't have any of it. It generates code in an afternoon that maybe took a team previously weeks to build, but none of those painful lessons come along for the ride. We dig into why 10x more code might mean 10x more failures. The concept of reliability guardrails, think ethical guardrails, but for keeping your systems up. Why you still have to test in production no matter how good your staging environment is? How Gremlin is rethinking their product for the world where agents, not engineers, are essentially the primary users. And why we're entering a painful, narrow part of the hourglass before AI gets good enough to handle all of this on its own. Watch On YouTube: piped.video/watch?v=Gbm8gaoO… @KoltonAndrus | @seanfalconer | @alexbdebrie

Making Data Agent Ready with Andre Elizondo Today we are talking with Andre Elizondo, the Head of Innovation at @mezmodata about their open source agentic harness for SREs called AURA. Mezmo got their start handling observability data at scale. Logs, traces, metrics, the usual stuff. AURA is their answer to a growing problem, as system complexity outpaces humans' ability to make sense of all that data, how do you actually make it actionable for AI agents? We get into their approach to context engineering, essentially making data agent ready before it hits the model. Why they built their own orchestrator in Rust? How they handle memory and self-correction in agent loops? Their take on MCP and where it fits versus Skills and code sandboxing and how the SRE role is evolving as agents become trusted teammates. Watch on YouTube: piped.video/vr6OA8ietsc @seanfalconer | @alexbdebrie

“You can’t test what you can’t see.” 👀 Modern AppSec starts with visibility. StackHawk maps your APIs from code → runtime → risk. See it. Test it. Secure it. 🦅 🎥 Watch the full interview to see how StackHawk is redefining AppSec. #AppSec #DevOpsSecurity #apisecurity

Modern Application Security and AI with Payton O'Neal This episode features a conversation with Payton O'Neal from @StackHawk, focusing on the evolution of application security and the role of Dynamic Application Security Testing (DAST). Payton argues that the shift from traditional security methods to more integrated, developer-friendly approaches is the way to go. The conversation covers the challenges and opportunities presented by AI in AppSec, the importance of bridging the gap between security and development teams, and the evolving threat landscape. They also explore the impact of AI on software development and the potential for AI to enhance security testing and vulnerability management. Watch On YouTube: piped.video/1exalnuYtno

Kentik CPO, @MavTurner, talks about everything from network security to the future of #AI and Kentik's new Traffic Costs solution in an interview with @TheSoftwareWith. Watch the interview: piped.video/watch?v=T9VGkrZY… #NetOps #NetworkSecurity #AIOps

Real-Time Network Cost Intelligence with Mav Turner The interview with Mav Turner, Chief Product Officer at @kentikinc, examines the evolution of network security and the critical role of network intelligence. Mav shares insights from his career and discusses how Kentik's platform provides visibility into network traffic to optimise performance and cost. The conversation also explores the impact of AI and machine learning on network management and the future of data centre operations. @mavturner | @JordiMonPMM | @TheSoftwareWith

Epic Web has something really big coming in coming days and Epic AI does as well. Stay tuned!

🚨 New from @GremlinInc: Reliability Intelligence is here. Now every engineer across your organization has access to the expert knowledge needed to run reliability tests, track down root causes, and fix issues quickly. Read more here: gremlin.com/blog/reliability…

In this conversation with Ben Lerner, CEO of @espresso_ai, we explore how Espresso uses Machine Learning to uncover optimization opportunities that traditional systems overlook—helping teams get significantly more out of their Snowflake spend. @ben_lern @JordiMonPMM

Espresso AI Renovates Snowflake Warehouses with Kubernetes techtimes.com/articles/31154…