Mistral re-enters the chat. Another strong open model will be available on @huggingface at the end of the month!

Agents test code, agents test agents. From one point of view, OpenAI could have guarded agents from "trying too much". From another one, wouldn't it bring the same decision fatigue as the "can I use this tool" request? Curious to hear how do you guys test your agents for misbehaving. ifixai.ai/breaking-news/open…

When I was an Android dev, I installed ten apps a day to better understand mobile app patterns. Today I work on an agentic product, so I play with a lot of agentic tools. Finally, I got my hands on Muse from Meta. It feels like an OpenClaw stripped down of all the hardcore setup. Too early for a deeper feedback yet, but the UI is really simple, and that's great. Referral codes time! Redeem my code in Settings within 48 hours of joining and we'll both get 1 billion Muse tokens. :) Code: G1YSWL muse.ai/join

There must have been someone who wrote a tool that creates user documentation from tests, right? Any recommendations?

Crazy times. Just half year ago I consulted a company about AI code reviews. We touched the topic of PRs and measuring productivity through PRs. PR count should never be the goal. But it was one of the signals we looked at. I shared that our team of about 30 engineers was shipping about 150 prs a week. The client told that it's about one-two things that gets done a day on average, and that it was not bad. Today, my team has 3 engineers - and we ship nearly 3 times as many PRs. New features and improvements get to our customers daily. And we work hard setting up processes and agentic loops and guardrails to increase the pace even further. If you’re not accelerating, you’re falling behind.

Ooops. That sinking feeling when you planned to spend a $15k grant of @AnthropicAI credits over the next 6 months of your new startup, and got paged that the credits have just expired 😅

Warp just went open-source - and it's been my favorite terminal app for three years running. The community has been asking for features like local LLM support for a while. With the codebase open, I'm optimistic those land sooner rather than later. warp.dev/blog/warp-is-now-op…

At Squire, we use AI coding editors, agents, code reviewers and UI test writers - the top-of-the-mind things. What else should we use to increase our velocity?

Software Engineers expected to use Cursor on interviews. I was a proponent of the approach for long time: having realistic interview experience. Why rotate a tree on an interview if you apply to build a web app for 3k b2b users? Why no AI tools when we use them daily?

Great example of catching the trend! A nice free tier + easy access from the Vibe Coding tools = success.

Unlock the secret to mastering Android App dev with #Kotlin! Just reviewed @PacktPublishing's 'How to Build Android Apps with Kotlin'. Love how it guides you from project setup to Store launch, enriched with hands-on projects. Perfect for beginners! #AndroidDevelopment

So, how does that Twitter thing work? Let me try... "As a team leader, I'm excited to hear about a better way of helping people in their career progression!" Is it done like that? :) #kcdc2023 #conference

How to get another bout of FOMO? Just read the list of this workshops at #kcdc2023! 😱 If you find something interesting, use promo code FriendOfArturBadretdinov to get a discount for the ticket:) kcdc.info/workshops