Thuru

😆

AI is gonna take your job and your girl.

🚨 Anthropic just dropped its 🦞 @OpenClaw competitor Meet Dispatch. A new research preview in Claude Cowork that completely changes how you interact with AI. Here’s how it works: 1️⃣ Pairs your phone to a persistent Claude session on your desktop 2️⃣ Message tasks on the go, come back to finished work 3️⃣ Executes code in a secure, local sandbox Your files stay 100% local and private, and Claude asks for your approval before touching anything Sure, the desktop needs to stay on, but the flexibility is insane. Rolling out now to Max users (Pro coming soon). Time to pair that phone! 👀

GPT-5.4 mini is now generally available for GitHub Copilot. github.blog/changelog/2026…

"Microsoft runs on Java. We have over 2.5 million JVMs in production across Microsoft" @JavaOne keynote!!!

I have been heavily using and pushing AI in engineering processes. In Aventude’s line of business, I get first-hand experience working with different teams, tech stacks, and budgets. Here are my learnings from the last 15–18 months. In the last 6 months, the relevance of code produced by AI agents has increased greatly. However, teams still struggle to get agents working well in projects, especially as complexity grows. When things are going well, a few bad specs can ruin the entire foundation. Learning #1: Have an .md file that tells the agent what it should NOT do. This covers not just syntax and referencing, but also the architectural essence to follow and how you want the code written. Learning #2: Beyond instruction-based coding, aiming for good results with fully autonomous agents seems like a very bad idea—especially since the inner loop burns money. Learning #3: Good specs lead to good code, fewer reviews, and so on. This shifts engineering effort toward the spec—spending more time refining it. I agree, but this cannot be dictated everywhere. Engineering the spec is a human task that has historically proven hard. No matter how well you engineer it, the spec will change. It becomes a continuous effort that eventually yields little net benefit. Agents writing 1M lines to create operating systems—these social-media stories are fine. But in business applications, code volume doesn’t matter; it’s about domain business logic, constraints, etc. Bringing that system-level taste to AI is hard. Overall, I see a push (mostly created by hype) to use AI. If an engineer feels they need to get hands-on—even if it takes a few hours more—doing so greatly benefits the long-term health of the system. Humans having an overall understanding of how the system works is a must. LLM logic cannot truly understand things; it tries to predict correctly. Understanding a system versus predicting its behavior are completely different.

This is so true.

🤯BREAKING: Alibaba just proved that AI Coding isn't taking your job, it's just writing the legacy code that will keep you employed fixing it for the next decade. 🤣 Passing a coding test once is easy. Maintaining that code for 8 months without it exploding? Apparently, it’s nearly impossible for AI. Alibaba tested 18 AI agents on 100 real codebases over 233-day cycles. They didn't just look for "quick fixes"—they looked for long-term survival. The results were a bloodbath: 75% of models broke previously working code during maintenance. Only Claude Opus 4.5/4.6 maintained a >50% zero-regression rate. Every other model accumulated technical debt that compounded until the codebase collapsed. We’ve been using "snapshot" benchmarks like HumanEval that only ask "Does it work right now?" The new SWE-CI benchmark asks: "Does it still work after 8 months of evolution?" Most AI agents are "Quick-Fix Artists." They write brittle code that passes tests today but becomes a maintenance nightmare tomorrow. They aren't building software; they're building a house of cards. The narrative just got honest: Most models can write code. Almost none can maintain it.

To this date I haven’t used minimal APIs in production. (Main reason , I haven’t started any project from scratch lately) but I wonder would I ever use it….. controllers are not complex IMO. Minimal syntax is cool though…….

@Devinbuild @hacking_newlife Research 🤣

@hacking_newlife Setting up a full time marketing agent that does research while I sleep

I was insanely skeptical on OpenClaw 🦞 Spent the entire day setting it up and I’m already hooked. Hired my first AI agents today • Nova (Chief of Staff) • Jarvis (Lead Engineer) • Sage (Social Media Manager) I now have an AI company working for me 24/7

We are still in the very early stages of real AI use cases in the world. Ubers and Airbnbs of the internet era took time to emerge; likewise, the Ubers and Airbnbs of the AI era are yet to come.

It's Friday. Your timesheet is blank. Your brain is blank. Your motivation is blank. Agent Mode in Edge for Business is coming soon to help take care of that last task so you’re one step closer to happy hour. Learn more: msft.it/6017QgBoF

We trained a new flood forecasting model designed to predict flash floods in urban areas up to 24 hours in advance. To help address a flash floods data gap, we created Groundsource: a new AI methodology using Gemini to identify 2.6M+ historical events across 150+ countries. We’re open-sourcing this dataset to advance global research, and urban flash flood forecasts are live now in Flood Hub to help communities stay safe.

Groundsource is a new AI-powered methodology that transforms millions of public reports into a high-quality record of historical disaster data to aid crisis prediction — starting with flash floods in urban areas.

@mkristensen Another question (actually a ceo asked me) : since ai is writing the code now, do we really need to worry which language it writes? Cant we hire anyone who’s available. Even if you want to change later you can use ai isn’t it?

When AI is writing more of your code, how important are code formatting and .editorconfig files to you?

people who dumped the money in the early copilot will agree to this.

Porsche reported its 2025 numbers and they are genuinely shocking. Operating profit came in at €90 million. The year before, it was €5.3 billion, and that is a 98% collapse. One of the most profitable car companies on the planet is now barely breaking even. Let me walk you through how this happened. In Q3 alone, Porsche posted a €967 million operating loss. That's over a billion dollars evaporated in 90 days. The full year was even worse than analysts expected. So how does the most prestigious sports car brand on Earth go from 14% operating margins to essentially zero in twelve months? It came down to three forces hitting at the same time, and Porsche had no answer for any of them. China was Porsche's golden market for years, the place where wealthy buyers couldn't get enough of the brand. That's over for now. Sales there collapsed 26% as local Chinese EV companies flooded the luxury segment with faster, cheaper alternatives that actually impressed buyers. Turns out the badge stopped mattering when the competition got that good. Then came the tariffs. The US hit European automakers with 15% import duties, and for Porsche, that translated to roughly €700 million in added costs over one year. You can't absorb that kind of hit when your volumes are already shrinking. But the biggest wound was self inflicted, and it's the one that should concern investors the most. Porsche bet billions on going fully electric and then EV demand across the industry stalled out. So they reversed course, scrapped their battery production plans, and decided to keep combustion engines around longer than expected. The cost of that strategic U-turn was €2.7 billion in write-downs in a single year. When you add up the restructuring charges, the tariff hit, and the EV reversal, total strategic costs hit €3.1 billion in 2025. Meanwhile they delivered 10% fewer cars globally, revenue dropped, and they're still paying for factory capacity they'll never fully use. Everything went wrong at once. The fallout is already in motion and Porsche is cutting 3,900 jobs by 2029 Internal documents suggest up to a quarter of the German workforce could eventually be let go, which would make this the largest round of layoffs in the company's history. The stock has lost more than a third of its value over the past twelve months. Here's why this matters beyond Porsche. If the most profitable automaker per vehicle on the planet can lose 98% of its operating profit in a single year, then no legacy car company is safe from the combination of Chinese EV competition, trade wars and a botched electrification transition. The auto industry is being rewritten in real time and the companies that hesitated on which direction to go are now paying the full price for that indecision.

AI has definitely fractured the software engineering process. In fact, it is a much-needed fracture! The difference between programming and software engineering is all about how a good engineer can design and think for the product. The product thinking and the architectural intuition always make the difference, now the truth is more open and augmented!

Junior and mid-level engineers can no longer push AI-assisted code without a senior signing off at AWS

Want to save $15-$25? Devin Review is a completely free PR review tool, with no signup required. Devin Review also supports: • Autofix • Smart diff organization • Copy and move detection • Codebase-aware chat Just swap github with devinreview on any PR to get started ⬇️