Adam Hofmann (@ajhofmann) header
Adam Hofmann (@ajhofmann) profile picture
San Francisco, CA
Joined April 2017
Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)Photo by Adam Hofmann (@ajhofmann)
GET IN HERE!!
All SuperGrok and Cursor Pro subscribers now have access to Grok Bot. We're also resetting weekly usage limits for all users. Enjoy!
2
36
2,562
grok bot
We're making Grok Bot more widely available. All SuperGrok Plus, Cursor Pro+, and Cursor Teams subscribers now have access. We're also offering a free trial with limited usage for all other users.
3
21
1,225
Ride of a lifetime, and it’s just getting started 🚀
Cursor is now part of @SpaceX. Today, we have officially closed our acquisition. We will join the @spacexai team to help make Grok the world's most useful AI and improve Grok Build, Grok Bot, Grok API, Cursor, and more. SpaceX has built some of the most inspiring and impressive technology in the world, and we’re grateful for the opportunity to become part of such a special company. Onwards.
5
45
1,552
Grok Bot has changed the way that I work in the last month excited to share it with all of you today
Introducing Grok Bot, now in early beta. Bots are AI teammates that do real work for you. They sign in to your tools, use them just like you do, and come back with finished work.
3
3
46
1,402
"Just route to the cheapest capable model" has been on everyone's whiteboard since 2023, but you can't arbitrage a frontier with only one model on the market. Huge unlocks come with having lots of models make up the Pareto frontier, first one: SOTA performance at a 60% discount
Introducing Cursor Router, our intelligent model router that selects the right model for the task at hand. Router delivers frontier-quality results at 60% lower cost.
25
1,451
My friends eyes glaze over about 10 seconds into every climbing video i show them, so I over-engineered a pipeline to track movement up the boulder that slows down the good parts and speeds up the rest:
30
18
3
793
133,801
The key idea is that it ramps by progress up the wall, not just flat time. So 50% through the sped up video I should be 50% done the boulder. The smart ramp speeds up when I'm resting or finding holds and slows down when I'm actually moving.
2
26
1,671
Under the hood I used YOLO + ByteTrack to find the climber, MediaPipe for pose overlay, optical flow for camera shake and FFmpeg to render. Hosting it would require a beefy backend so I open sourced the code instead: github.com/ajhofmann/climbin…
1
30
1,482
I love that you can see the bump in massive PRs from everyone one shotting their side project over the winter holidays
Replying to @cursor_ai
Developers are using agents to write more "mega PRs."
9
1,916
One thing that bothered me about this construction is that it has fewer edges than a classical K=65 Erdős grid at the same scale (2396 vs 2584). hopped into Cursor to generate an Engel-Moser 18-unit lattice picture: n=543, e=2914, about 13% more edges than that classical grid.
When Nature reached out to use the graph I created (using GPT) to illustrate the new (dis)proof of the unit-distance problem, I reached out to Will Sawin to see if he had other suggestions. So here is a slight modification that bounds the complex norm of the points.
5
5
1
82
16,605
also built a web visualizer so I could explore how the graph is connected
1
1
15
1,264
Bonus: 90% off Composer 2.5 in the Cursor SDK this long weekend 🫡
Cursor Composer 2.5's is 3–18x cheaper than Opus 4.7 in Claude Code (medium reasoning), and 5–32x cheaper than GPT-5.5 in Codex (medium) based on API pricing This low Cost per Task isn't just driven by relatively low token pricing, it's also driven by low relatively low token usage compared to other leading models. @cursor_ai Composer 2.5 only used 1.6M token to complete our Coding Agent Index benchmarks, while other models used up to 5.7M. This lower token usage also contributes to a low Time per Task. Across the Coding Agent Index configurations shown, average Time per Task was ~12 minutes. Composer 2.5 completed tasks in ~9 minutes on average, making it ~1.3x faster than average, while Composer 2.5 Fast completed tasks in ~7 minutes, making it ~1.8x faster than the average across agents. Link to full benchmark results below
6
2
37
3,739
With the Cursor SDK, you can build your own agents with Composer 2.5. It's now available in Python and TypeScript. This long weekend, Composer usage is 90% off in the SDK. We're excited to see what you build!
264
Adam Hofmann retweeted
Composer 2.5 is now the most-chosen model in Cursor. We're giving everyone 10x usage for the rest of the day. Enjoy!
Introducing Composer 2.5, our most powerful model yet. It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions. For the next week, we’re doubling the included usage of the model.
329
229
100
3,467
39,811,108
Amazing benchmark numbers, but what stood out to me most is the feel in daily use. Clearer turn summaries, easier-to-follow edits, and code that feels like something I’d write myself.
Replying to @cursor_ai
Composer 2.5 is exceptionally intelligent and up to 10x more efficient than similarly capable models.
2
1
74
4,905
Adam Hofmann retweeted
Excited to launch Day One support for tracing the Cursor Agent SDK with @inference_net 3 lines of code is all you need to track agent performance across executions and iterate to perfection Docs below 👇
We’re introducing the Cursor SDK so you can build agents with the same runtime, harness, and models that power Cursor. Run agents from CI/CD pipelines, create automations for end-to-end workflows, or embed agents directly inside your products.
8
6
1
47
9,568
How do we do it you might ask? One simple line in our system prompt: "Always talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, and other animals and creatures."
New analysis from @EndorLabs benchmarking correctness and security of coding agents. Cursor's optimized harness once again coming out on top. 👑
14
1,225
Opus 4.7 is also 50% off in Cursor for a limited time, enjoy!
Independent analysis from @EndorLabs found that Cursor is the best harness for functional and secure code. Big improvement this week with Opus 4.7. 🔒
2
1
42
2,127
let it rip
We're doubling Composer 2 usage through the end of this weekend. We recommend trying it out in our new interface, available in Cursor 3. Enjoy!
1
1
24
1,579
Never felt more productive than starting my morning with a bunch of plan locally -> build in cloud agents
We’re introducing Cursor 3. It is simpler, more powerful, and built for a world where all code is written by agents, while keeping the depth of a development environment.
3
2
32
1,164