Social Wire
Page 13 of 49 · newest first.
@russelljkaplan · Cognition
The software engineering landscape is becoming increasingly 'intelligence saturated,' where cost and speed are now the key differentiators for AI models@ScarletKc Ambassador · Cloud
Ambassador ScarletKc compares Devin Review's auto-review + auto-fix favorably to CodeRabbit and Codex — automated PR review with fix capability, not just comments.@Toolglade · Desktop
Choose the right coding tool for your needs: Cursor codes with you, while Windsurf (now Devin Desktop) codes for you, both available for $20/month@ScarletKc Ambassador · Cloud
Chinese community breakdown: Devin Fusion scored 68.1 on Artificial Analysis — ~2 points above Claude Code with Opus, 31% cheaper per task. Multi-model orchestration wins on both metrics.@vstalingrady · CLI
Optimize your costs: a recent comparison of Devin CLI and Gray plugin pricing reveals that Gray sent fewer input tokens per request on 7 out of 8 tasks, making it a more cost-effective option for some users@realmtbman Ambassador · Cognition
Ambassador Yohann Calpu reports: packed Cognition event at University of Toronto. Student developer interest in AI coding agents remains strong.@willebrew Ambassador · Cloud
Ambassador Will Killebrew highlights how Devin's auto-approvals + Coach combo balances velocity with safety for teams mixing experienced and newer engineers.@aitrackerbot · Desktop
The latest Devin Desktop (Windsurf) Next build confirms the addition of Theia, Obsidian 2, Obsidian 3, and Slate 2 model codenames, each with its own reasoning levels, though the lab behind them and their release dates remain unannounced.@fayequepeerzade · Desktop
Four new model codenames - Theia, Obsidian 2, Obsidian 3, and Slate 2 - have been discovered in the latest Devin Desktop Next build, hinting at potential future model releases or updates.@noah_z_zhang · Cognition
Users are praising the sleek design of Devin Next, the beta version of Devin, highlighting the continuous improvement in user experience.@razmichaels · Cognition
DevinAI's voice mode is impressing users with its human-like voice inflections and humming, showcasing the advancements in AI-powered communication tools.@marclou · Cognition
Join the 13th Indie Hackers meetup in Dubai on October 25, organized by Marc Lou and supported by Cognition, to connect with fellow developers and learn about the latest advancements in the field.@suprmind_ai · CLI
A recent comparison highlighted the efficiency of Argon in Devin CLI, completing a task in 13 minutes for $1.2, compared to Opus 5.5 in Claude Code, which took 15 minutes and cost $8.2.@alexhvnsen · Cloud
Codex and DevinAI sub agents are being used together to create a powerful loop, with impressive results and potential for further innovation@alexhvnsen · Cloud
Independent devs are building and open-sourcing tools to make Hermes agents better than Grok bot, with a focus on simplicity, speed, and convenience@Nicolo_Tognoni · Cloud
Devin Oncall is a game-changing agent that takes production alerts to the next level, finding root causes and opening fixes for review@noah_z_zhang · Cloud
Upgrading to Devin Max can unlock unlimited potential, allowing users to cook without limits and take their productivity to the next level@Avi31m · Cognition
Join the DevinAI meetup in Dubai on October 25th, featuring demos, credits, and more@codeitlikemiley · CLI
Devin CLI shows promising performance, with no jank when scrolling compared to Claude Code@ekcheungAI · CLI
Devin CLI's Argon model outperforms Opus 5.5 in 3D build tests, with faster completion and lower cost@dzienko Ambassador · Desktop
Get ready for the new Devin Desktop, launching next week with exciting updates and improvements@ItsCynnamoroll · Cognition
DevinAI is pushing the boundaries of innovation, refusing to settle for anything less than world-class solutions and engineering their own from scratch to create the best agent-native git service@cognition · Cognition
Cognition and ND Studio launched American Forge—a six-month residency that embeds engineers inside government agencies to build software where it's needed most.@devdudejeremy · CLI
Heads up: SWE-2's free access ends today, with pricing set at $3 in and $15 out per 1M tokens afterwards, according to Cognition's pricing page and models doc