max retweeted
"show me software with 0 bugs and ill show you software with 0 users"
The design team killed it. Go to a Captain and swipe left… I’m in love with this flow.
Well done @capydotai 👏
18
Alright, time to make a startup. Who’s in?
Lunch is on me!
✨ Introducing DevinDash ✨
Starting today, dash food with the DoorDash MCP and a Devin Bot.
Startups can claim $1k of free DoorDash: devin.ai/startups
1
46
If you haven’t yet, you need to give @Da7_Tech a follow.
He puts a ton of effort into his content, no frills, no slop, and entirely honest feedback.
Introducing Da7em Bench v2.0.
A fully independent benchmark for AI models, built on real client work. The first of its kind in the world.
How it works
I test each model across a wide range of scenarios. I compare its decisions with the choices people make in the same situations and with what people generally agree is the most sensible choice.
Models still fall behind humans in several areas. A higher score means the model’s decisions are closer to human choices. A lower score means they are further away.
As before, I test each model impartially across multiple harnesses. The standard is professional work. I judge each result against what a paid professional would deliver for the same brief.
Version 2 covers 11 models across 12 axes. Each axis is scored out of 10, and the overall score is the average of all 12.
The tasks and evaluation methods stay private so they don’t leak into training data and affect future results.
Da7em Bench helps you choose a model for each task. The model with the best visual taste may not be the best writer. The model that plans best may not be the best at carrying out that plan.
Da7em Bench is fully independent. No sponsors or outside funding. I’ve paid for everything out of my own pocket, thousands of dollars so far. There are many more models I’d like to include. This is a personal benchmark, and covering them all on my own takes time and work.
The four images show the overall ranking, the scores for every model on every axis, the twelve axes grouped into four pillars, and a side-by-side comparison of Claude Fable 5.2 and GPT 6.2 Astra.
You can explore the full benchmark and compare any two models here.
da7tech.com/da7em-bench/?lan…
I’m seeing this a lot on the feed…
Let’s stop pretending that it’s hard to remember which model is which. Plus, every UI will have them ordered for the non-X addicted user. Google cooked, let them take their victory lap.
Nothing against Kun, this quote is purely an example.
try and answer this without looking them up -
gemini 4 argon, hydrogen, neon, carbon
which is smaller, which is bigger?
15
Devin Bots are so useful that I’ve been running them non-stop since release. I’ll never get over how nice texting a bot and receiving a full QA video of an app is.
@cognition consistently launches iterations of these types of products that are actually useful for developers.
We’re giving away up to $100M in Devin Bots this week!
Just text Poke and convince them to build your idea:
poke.com/bots
26
max retweeted
i miss coding before AI like i miss gaming all day as a kid
it was a simpler time. and it was pure bliss.
I had a few microservices go down last night and had to do this exact thing.
I’ve never set up a new feature so fast in my life. Just wish I had it last night 😭😭😭
Introducing Devin Oncall: Devin automatically investigates alerts and bugs, helping your team diagnose and resolve production incidents.
Devin Oncall already handles thousands of alerts every day for some of the world’s largest companies. Let it become your SRE today.
41
People underestimate how big of a deal this is - well done @cognition 👏
max retweeted
Cognition is giving away 50 $200 Devin Max plans to celebrate the new model launches! ⚡
Now available in Devin:
• GPT-6 Astra, Sol, Luna + others
• Claude Opus 5.5, Fable 5.1 + others
• SWE-2 (Free until October 15)
• Fusion Frontier harness (Fable, Astra, Sol, Opus)
• Gemini 3.8 Flash + others
• Grok 4.7 + others
• Kimi K3 + others
• Inkling
• DeepSeek V4.1 Flash + others
• GLM-5.3 Flash + others
• Cloud agents on Linux, macOS, and Windows
To be eligible, reply below with what you're building (or something you'd like to build with Devin)! We will choose winners in 24 hours.
max retweeted
/goal is now my default mode of working with agents. but I spend a ton of time preparing all of the context and rubric criteria before kicking off the job
imo every agent message that’s not attached to a /goal should be in service of generating a spec + rubric to launch with /goal
because:
1. “verifiability” of a dispatched sub-task is the largest determinant of an agent successfully completing it
2. “complete this task to completion” is everyone’s desired UX, no one wants a half done solution, that’s what /goal tries to give you
3. thinking in terms of verification and units of work is a good mental model for thinking about sending agents to do useful things
















