‼️it seems Anthropic is ready to publicly launch a new version of Mythos, something better than Mythos Preview.
a codenamed model “Oceanus” was given access to some red teamers yesterday according to @synthwavedd.
it’s apparently been paused already, due to someone reselling access through a Chinese API proxy lmao 💀
Mythos pricing might also end up at with
$16 Input, $80 Output according to @scaling01
yup i basically uninstalled every other tool and just kept devin desktop now.
Why not zed? cause they dont support devin which been my main agent. Also devin-desktop ui is cleaner imo
Really proud of the @cognition team for launching Devin Desktop today. Just remember you still get all the features you had with Windsurf, but now you can add in your other agents as well.
When you launch Devins from your other surfaces (Slack, API, Devin Automations, etc) you will see them pop up in Devin Desktop as well. Where you can check the status of your fleet of agents and unblock things where necessary.
Introducing Devin Desktop: the next generation of Windsurf
Manage fleets of local and cloud agents from one surface
Support for any ACP-compatible agent
With a full IDE for when you need to jump into the code
Introducing Devin Desktop: the next generation of Windsurf
Manage fleets of local and cloud agents from one surface
Support for any ACP-compatible agent
With a full IDE for when you need to jump into the code
Yup, tried it yesterday and its much better than old cascade imo. MCPs and custom slash commands work so u can switch if u dont rely on the other features
Cascade is the previous local agent powering Windsurf.
Devin Local is our next generation harness, shared with Devin CLI.
We recommend switching to Devin Local going forward.
It's still in preview, so a few Cascade features aren't supported yet: Memories (context the agent auto-saves and recalls across conversations), Workflows (reusable multi-step prompts triggered via slash commands), and Codemaps (visual maps of your codebase's structure).
Unpopular opinion: using this model to "reduce cost" is a cheap mode.
Codex sub gives you kinda unlimited GPT 5-5 state-of-the-art models for $200, why in the world would you use this for coding?
AMD had quite the announcements
• RX 9070 GRE announced
• Starts at $549
• 12GB VRAM
• 48 compute units
• 48 ray tracing accelerators
• 96 AI accelerators
• 220W board power
• Ryzen 7 5800X3D 10th Anniversary Edition is REAL
• Ryzen 7 7700X3D coming to AM5
• AM5 support confirmed through 2029
• Ryzen AI Max PRO 400 with up to 192GB memory
AM5 through 2029? That will keep some smiles.