Social Wire

Page 27 of 49 · newest first.

@valenxi · Desktop
Devin Desktop's Fusion model is helping users get more out of expensive models like Fable 5.1, making it a valuable tool for those looking to optimize their workflow
@Chris_Wozniczek Ambassador · Desktop
GPT-6 Sol, GPT-6 Luna, and Claude Opus 5.5 are all live in Devin Desktop and CLI — and Chris Wozniczek is comparing them against SWE-2 on FrontierCode 1.1 Extended. For teams benchmarking locally, all three new models are available to route through Desktop and CLI now.
@yellowfox_h4xor · Cloud
Devin's cloud VM and SSH capabilities are providing users with a unique and powerful tool for steering and controlling their workflow, making it an attractive option for those looking for a more efficient and effective way to work
@zenoware Ambassador · Cloud
Devin's MCP server for security checks works across any harness — Codex, Claude Desktop, Amp, OpenCode. A practical example of how Devin's interoperability layer lets teams route security workflows through their existing tool chain.
@blueemi99 · Desktop
User praises Cognition and Devin for their innovative features, including cloud agents, free models, and Fusion, but suggests improving the desktop app.
@bryann2k_dev · Cognition
Get unlimited access to Cognition's new SWE-2 model, built on Kimi K3 and tuned for software engineering, for just $20 until October 10.
@cognition · Cloud
Devin Cloud now runs directly in your terminal with `/cloud` commands and `devin ssh` for full VM access—create, steer, and resume sessions from the CLI, then `/handoff` work back to your local machine without leaving the command line.
@Chris_Wozniczek Ambassador · CLI
Get insights into your Devin token usage with the new Otter Stats app, which provides detailed statistics on token burn, project consumption, and more.
@cognition · Cloud
Grok 4.7 joins Devin's model lineup with a 59.4% FrontierCode 1.1 Extended score—xAI says it excels at hard backend engineering work, now available in Devin Cloud for complex system tasks.
@Da7_Tech Ambassador · Cloud
Ambassador @Da7_Tech reports Astra (Devin Desktop's preview model) completes tasks in ~15 minutes at light effort—a fast, cost-effective choice for straightforward coding work that doesn't need frontier reasoning.
@devindesktop · Desktop
Grok 4.7 is now live in Devin Desktop and Devin CLI—xAI's latest reasoning model ready for local agent workflows and terminal sessions.
@frmdomi Ambassador · Desktop
Devin Desktop Cloud now prompts users for missing credentials, allowing for safe submission without exposing them in the chat.
@ikavalerov · CLI
User reports discrepancies between Devin Desktop and CLI, specifically with the handoff feature, and seeks resolution.
@jumperz Ambassador · Cognition
The team is working on restocking StepFun plans after high demand, with plans to make them available again soon.
@kinopee_ai Ambassador · CLI
Devin CLI now connects directly to Devin's Cloud VM from the terminal, letting you send instructions and switch repos mid-chat with multi-select—no GUI needed for full Cloud session control.
@kinopee_ai Ambassador · Desktop
In Devin Desktop, the keyword "megaplan" triggers deeper reasoning during plan creation—an undocumented prompt modifier Ambassador @kinopee_ai uses to get more thorough multi-step strategies before execution.
@kinopee_ai Ambassador · Cloud
Devin Cloud sleeping sessions now resume the instant you start typing—the VM wakes in the background so it's warm and ready by the time you hit send, cutting wait time to zero.
@LucasDolutech · Cloud
User prefers Cognition's cloud environment over others, citing excellent performance with Devin's Max plan and SWE-2.
@nick_white Ambassador · Desktop
Consider Devin Desktop from Cognition as a top alternative to other IDEs, offering a local editor and cloud agents for a unique development experience.
@notjazii Ambassador · Cloud
Ambassador @notjazii ran the same task on max reasoning—Grok 4.7 finished in 32 min at $8.14 while SWE-2 completed it in 15 min at near-zero cost, showcasing the tradeoff between frontier reasoning and fast preview models.
@ScottWu46 · Cloud
Ambassador @ScottWu46 highlights a deep dive from @jkelleyrtp on Devin Mac—check the original post for the full breakdown of how Devin's local Mac agent works under the hood.
@szavinny · Cloud
Don't miss out on the exceptional SWE 2.0 model from Cognition, available for free this month, and pair it with Fable 5.1 for the ultimate setup in Fusion mode.
@YarBirzool · Cognition
User reports payment issues with Devin using a Ukrainian card, despite success with other AI services. The Cognition team is encouraged to investigate.
@b_shulha · Cloud
Cognition's Fusion mode has impressed users with its ability to optimize costs, with one user describing it as reaching a 'WOW level' in terms of efficiency and future potential.