
I believe in Christ. Also I'm a web developer who's learning design thinking. ex-@neoscms team member. #reactjs all things! In love with #opensource
WFH
Joined January 2011
- Tweets15.5K
- Following1.7K
- Followers828
- Likes15.2K
Dmitri Pisarev retweeted
If your company needs software dev help, I’d love to talk to you! I have some amazing experienced US-based devs on the bench who are very good at using agents and have a high quality bar.
We are best at mobile and web and desktop apps, but can do much more than that.
Average experience level is 15 years. You’ll love working with them. Great people, excellent devs.
Appreciate retweets and any referrals!
Dmitri Pisarev retweeted
When I try to roll the windows down in my Roadster and I accidentally launch the car 60 feet into the air
Dmitri Pisarev retweeted
Open Al's internal benchmarks show GPT-6.1 Sol crushing Opus 5.5, with Anthropic struggling to keep up
Dmitri Pisarev retweeted
You'll know when AI gets de novo reasoning ('AGI'), because it won't look anything like 10,000 monkeys exploring 991,235 dead ends before finding the correct path.
Dmitri Pisarev retweeted
Replying to @DavidF1344 @josephflaherty
My grandparents and great grandparents could not fathom why some Boomers and Xers wanted to move back to the farm. It wasn't something they romanticized, it was a hard reality they couldn't wait to leave. Being close to nature means being acquainted with her teeth
Dmitri Pisarev retweeted
Replying to @dimaip @code_barbarian
You and me both
Believe it or not, I went through a "I don't want to use Mongoose anymore" phase about 10 years ago
Then I found myself re-implementing everything Mongoose already did; wondered "why am I doing this to myself?"; never looked back
Dmitri Pisarev retweeted
🌟 What's New in @mongoosejs 9.9: Major Performance Improvements
thecodebarbarian.com/mongoos…
Dmitri Pisarev retweeted
I remember when prototyping something took days and shipping it took weeks
Now prototyping takes minutes and shipping it takes weeks
Dmitri Pisarev retweeted
Have you built a language model? You should. It's so much fun to chat with something you made.
Anyone can do it, too. I made an app that teaches the fundamentals and gives you everything you need to build your own: languagemodelbuilder.com
Dmitri Pisarev retweeted
argentina fans proving that even american cities are walkable if you're not lazy
Dmitri Pisarev retweeted
Co-sign ✍️
Agent harnesses need to implement proper tree-like threading that isn't an afterthought. Pi /tree is a good example.
Side-chats and /btw are ephemeral for no good reason, sub-agents lack interactivity, forking/branching is clunky and too much of a clean break.
Dear Codex team,
I am a huge supporter of your work. I really believe in the direction you are going, and I enjoy using Codex and the Codex app more than I can ever describe in words!
Now I will stop fanboying for a second, and I will ask you to please, please fix side chats. They are a huge part of my workflow, but there are several issues with it, like:
1. You can't fork messages from side chats.
2. Plans don't work.
3. Expiration removes side chats completely from history. I know they are ephemeral and that I can't send new messages after some time, and that's okay, but history is important. I need to be able to read some of these, hours or even days after they expired.
But, there is a new issue that has been happening since the second-to-last update, I think. It has the misconception that, when on a side chat, it is on a restricted sandbox that cannot even reach my services, even if I ask it to.
Obviously I have given full access. I always had. So I don't think I've done anything wrong here! Now, when that happens, I need to go to a new thread or the main thread and disrupt the whole flow.
So that's a request from your beloved user. I love you! You love me! Please fix it!
Byeeeeeee
Dmitri Pisarev retweeted
Everyone hyping loops right now is going to absolutely lose their minds once they learn about state machines
Dmitri Pisarev retweeted
Replying to @cnakazawa
Dmitri Pisarev retweeted
Too many developers don't understand what "compounding slop" is.
A loop that prompts agents is a great way to automate slop creation. Constrain the state-action space so the loop can't drift, then automate inside it.
Human-in-the-loop = feature, not bottleneck.























