@MattSchrage

building @cognition. prev: co-founder of @fig (acq by amazon) @ycombinator @foursquare @harvard

 San Francisco
schra.ge
Joined November 2011
  • Tweets3.5K
  • Following1.8K
  • Followers2.7K
297Media
i'll tell my kids that this is what an 'agent sandbox' was
Nico DiPlacido
@tweetNeeksOct 2
They sell these new things for devs that look like call booths so no one's suspicious, but really they're cages that lock from the outside, and its linked to GitHub so it won't let them out until they release their feature. It takes a little bit of getting used to, but it's super effective.
1
8
587
Matt Schrage retweeted
Brendan Falk
@BrendanFalkSep 25
Man, @fig was so ahead of its time
almonk
@almonkSep 25
I love the idea of complementing the user's shell. Augmenting their experience rather than attempting to hijack it. Our latest experiment allow you to insert a path to ✨anything, anywhere✨ with instant fuzzy matching as well as traditional path navigation. As useful for your `cat` as your `claude`.
20
2
1
188
16,810
sam
@samdapeSep 25
hard economic times
11
845
friends don't let friend run agents locally normanponte.io/19df691f
2
12
544
Historically, when working on a new feature, you've had to choose between a local agent - interactive with a human in the loop - and cloud agents, that have their own machine that can run for hours autonomously. I think we've finally nailed the execution of handing off back and forth between the two. Try it now by running /cloud from the CLI. devin.ai/blog/devin-cloud-in…

Bringing Devin Cloud to your terminal | Devin

Starting today, you can create, steer, resume, and watch your Devin Cloud sessions right from your favorite terminal. Just use devin --cloud or /cloud.

devin.ai
3
1
38
1,417
Devin is the only cloud agent on the market that lets you natively spin up macOS, Windows and Linux machines. How it felt flipping the feature flag:
Jon Kelley
@jkelleyrtpSep 15
Devin is the first cloud agent with macOS, Windows, and Linux . Nobody else has it because it’s really fckin hard to do. 🥲 We rebuilt storage, networking, VNC, and CUA from scratch - in Rust! Super proud to be working alongside the great minds of @cognition and @dioxuslabs
8
4
2
113
5,172
macOS sandboxes for cloud agents is such an obvious idea, but the reason why no one has done it yet at scale is because it is fiendishly difficult from an infra perspective and requires spelunking into macOS internals. Now you have no excuse: go build that app idea with Devin!
Cognition
@cognitionSep 15
Special delivery: Devin just got a Mac 🍎 Now Devin can: 1. Build & test apps on its own Mac VM with iOS simulator 2. Send a screen recording via Slack 3. Send a TestFlight link so you can start using it 📲
13
6
2
92
15,347
I've followed Dioxus since the @fig days, when we were looking for a Rust UI framework. @jkelleyrtp is the 🐐 - it's going to be a blast working with him, @demonthos & @nicoburns.
Jon Kelley
@jkelleyrtpSep 10
We have some very big news to share. 🎉 Today, Dioxus Labs is joining @Cognition to accelerate the development of Devin, Cognition’s autonomous cloud coding agent. We are excited to continue building Dioxus while also helping Cognition shape the future of software engineering.
1
13
1,146
love that new model smell, but seriously @silasalberti and the entire research team cooked with this! 🧑‍🍳
Cognition
@cognitionSep 10
Introducing SWE-2, our closest model yet to the frontier. On leading evals, it scores on par with recent frontier models – at up to 70% lower cost. We scaled RL to multiple trillions of parameters, with a refined recipe that pushes the Pareto curve on both capabilities & cost.
2
1
30
1,528
If there is one takeaway I have from my time at @cognition it's that 'the future is already here, it's just not evenly distributed'. Incredibly excited & humbled to be building the future of software engineering with such an amazing team.
Cognition
@cognitionSep 8
The world needs far more software than it can build. Cognition exists to change that. We’ve just raised over $2B at a $48B valuation, led by a16z, Accel, Founders Fund, General Catalyst, and Avenir. Since our round in May, run-rate revenue has grown from $492 M to almost $900 M.
6
54
2,074
This was a fun one! Building an agent that feels at home in Slack is a surprisingly hard problem - there is a lot of informal social etiquette, like when to tag people, use a DM, wait for someone else to respond, etc. I’m pretty excited about what we shipped! If you haven’t tried @DevinAI in Slack yet, now is a great time to check it out!
Marc Benioff
@BenioffAug 20
Don’t code alone. Slack Code is live. Humans and agents. Same channel. Same work. Launching today with agents from @AnthropicAI, @github, @Cognition, and @vercel. This is real multiplayer coding. See it at @Dreamforce #DF26
7
2
1
53
10,418
Read more about the process: devin.ai/blog/devins-slack-e…

Devin's Slack Etiquette | Devin

How we refined Devin's behavior in Slack over the past year, across prompts, Devin's harness, and our usage of the Slack API.

devin.ai
2
131
AGI has been achieved internally! @DevinAI is just like us and is now capable of getting distracted from its work by playing video games
3
5
1
39
10,387
fusion is fun to play with - not only is it cheaper, but it's often faster as well
Cognition
@cognitionJun 29
Conventional model routing sucks. It passes benchmarks but fails to write code you'd actually merge. Introducing Devin Fusion, a new hybrid-model harness for agentic coding. In testing, it reduces the cost of Fable-level intelligence by 35% and still feels good to use.
1
1
17
2,537
Matt Schrage retweeted
siddharth
@buildwithsidJun 13
devin cli > kiro cli > opencode cli > claude cli > codex cli > antigravity cli > any other based on my experience
27
2
1
105
8,756
gotta go fast
nader dabit
@dabit3May 20
Tokenspeed from from @mikeveerman is a nice and fun way to visualize token speeds. At 950 tokens per second, I paired it up with SWE-1.6 Fast from @cognition in Devin for Terminal to get a feel for the difference between pure code streaming and watching an agent work when the underlying model runs at 950 tokens per second. At this speed they are both blazing fast. mikeveerman.github.io/tokens…
1
10
1,681
I’m pretty sure that thanks to our partnership with @cerebras and our optimized agent harness, Devin CLI is the fastest coding agent available. 👇Try it out for today + if you send me useful feature requests / bug reports, I’ll give you free credits!
Harkirat Singh
@kirat_twMay 18
Devin cli is something. Feels snappier than the rest.
5
4
40
5,088
devin.ai/terminal

Devin CLI — The AI Software Engineer in Your Terminal

Devin runs locally in your terminal. Every frontier model, Fusion for frontier results at over a third less cost, and a hand-off to Devin Cloud when you close your laptop.

devin.ai
1
300
Load more