Social Wire

Page 52 of 53 · newest first.

@willebrew Ambassador · Cloud
Builders keep calling out Cloud’s autonomous test + video recap loop: you don’t have to babysit the browser; you get a replay of what Devin actually clicked. Same thread as “record a demo for me” — proof artifacts, not vibes.
@cognition · Cloud
From the Wire (@cognition): Ship straight from the conversation. Devin in @Slack lets your team investigate issues, answer codebase questions, and start dev tasks without leaving the channel. Thanks for the feature, @Slack👇️ Source: https://x.com/cognition/status/2077484850244648998
@dabit3 · CLI
From the Wire (@dabit3): Since you don't need a frontier model typing every line of code, we can optimize subagents beyond just cost. /lightning skill offloads implementation to 1,000 tok/s subagents, resulting in ~5x faster execution. it… Source: https://x.com/dabit3/status/2077456437652119693
@digitalapplied · Desktop
Handy roundup of what sellers are selling in the 2.2 era: faster boots, Review as a PR pre-pass, and “use Cursor + Claude Code + Devin by job shape.” Take the flashy computer-use claims with a grain of salt until you’ve watched a real session—but the “start with tightly scoped tickets” advice is still the durable part.
@ivanda_mislav · CLI
Fresh ops pattern: Devin CLI inside a Daytona sandbox — headless prompts, `--continue` memory, free-tier friendly. Good blueprint if you want CLI agents isolated from your laptop without standing up Cloud for every experiment.
@_markfenner Ambassador · Desktop
DevinX external TestFlight is live: pocket mission control for Cloud and local Mac sessions, with on-device dictation so prompts never leave the phone. If you’ve been waiting to steer Devin without sitting at the desk, this is the build to grab while tester spots last.
@developersdigest · Cloud
Parallel agents don’t erase review. After a 200+ PR fan-out day, Developers Digest’s punchline stuck: Devin Review becomes the queue when you tag it on every PR. Desk story:
@hn · Cognition
HN’s best thread on Review wasn’t “does AI catch bugs?”—it was “don’t let a faster review UI become an excuse to skip reading the diff.” Cognition’s own framing (organize so *you* understand the PR) is the right bar. More:
@albertgao · Cloud
Practical multi-agent desk pattern: Devin Cloud Fusion builds the ticket, Codex Desktop reviews, loop until the report is clean. Same lesson as our Review bottleneck story — fan-out needs a review lane, and people are wiring two products to get it.
@jjacky · Cognition
Sentiment flip worth tracking: ex-skeptic now pointing at Cognition’s bet on cloud agents, task insurance, and QA as the 6–12 month trajectory — not just another IDE autocomplete. Desk note, not a press release.
@_markfenner Ambassador · Cognition
From the Wire (@_markfenner): Let's truly test the new algorithm. X just started boosting mutuals so clusters of shared interest can form. L X, I want to connect with: Devin builders Hermes agent users Security-first shippers Vibe coders who… Source: https://x.com/_markfenner/status/2077452829896343950
@_markfenner Ambassador · Desktop
From the Wire (@_markfenner): 8 miles this morning🏃🏾‍♂️. Early access was ready by the time I got back. DevinX. @DevinAI in your pocket. Cloud and local sessions, one app. We really do build online and offline over here. 🦫 Who wants the link?! Source: https://x.com/_markfenner/status/2077396256549130387
@frontierai · Cloud
Worth re-reading as a scoping checklist, not a 2026 scorecard: Devin crushed well-defined ~1-hour fixes and choked on multi-component full-stack work. The lesson aged better than the ACU prices—narrow the ticket, teach codebase patterns, abandon loops early. Treat it as field archaeology plus durable advice.
@startdevin · CLI
Japanese desk tip from StartDevin: Devin for Terminal’s `/handoff` is the boring superpower—close the laptop, keep the job alive in cloud, pull it back locally later. We fold that into the Desktop cutover story:
@SlackHQ · Cloud
Slack’s own push: Devin in-channel so the request that never becomes a ticket still gets investigated and started without a context switch. Cloud agents earning the place where work already lives — not another tab.
@RDarrylR · CLI
AWS Lambda now offers a one-click setup for coding agents, streamlining the configuration process and supporting multiple agents, including Devin Desktop.
@the_vc_intern · CLI
Cognition's acquisition of Windsurf has led to significant growth, with a revenue run rate increase from $73M to over $500M in just 12 months.
@robinebers · Desktop
A user suggests synchronizing features and credit policies across Devin products to enhance user experience and consistency.
@thsottiaux · Cognition
From the Wire (@thsottiaux): Or… what if we gave you $100 in Codex credits if you tell us what you love about GPT-5.6 Sol or why you switched? Tweet it, claim your gift, enjoy more usage. First 10k get the free tokens! Source: https://x.com/thsottiaux/status/2077248807533003257
@vijayiyengar · CLI
Pinecone's creators built their own coding and knowledge work agent to unlock cloud-based productivity, ensure accessibility for all teams, and enable end-to-end testing.
@chatforest · Desktop
Windsurf’s June OTA wasn’t just a rename—Devin Desktop puts the agent Kanban first, and Cascade’s July runway (FAQ: available through July) forces the migrate to Devin Local (shared CLI harness, still preview). ACP support is the quieter signal: third-party agents can sit beside Devin in the Command Center. Full desk story:
@YouTube · Desktop
A practical deploy walkthrough that shows Devin Desktop moving from code to production — the kind of verifiable, end-to-end workflow Cognition's docs recommend as a sweet spot for cloud Devin.
@YouTube · Cognition
SWE-1.7 is getting international hands-on coverage. Cognition launched it as their most capable and efficient model to date; this review puts that claim in front of a Japanese developer audience — a market Cognition entered formally in April with Cognition Japan.
@_markfenner Ambassador · Desktop
Desktop operators are getting a real pocket model picker for local DevinX sessions—Sol, Fable 5, GLM, Kimi, or Adaptive, plus reasoning effort. That matches the wider “don’t hard-code one lab model” pitch we’ve been seeing on Cloud Fusion. Worth testing which default you actually leave on overnight sessions.