... citation in a popped-out chat and the actual email opens in the main window which politely un-minimizes itself from the tray instead of making you go find it. Open a second, third, fourth chat window they cascade instead of stacking pixel-perfectly on top of each other, which was funny exactly once. Fixed Follow-up questions stop re-searching for mail they already have. The digest marks everything it covered as read, then the follow-up went hunting ...

... about the things they didn't check.But I'd go further than just flagging what wasn't checked. I'd want it to default to naming its own confidence level per claim, something like "ran the tests, they passed" versus "assumed this worked based on the diff, didn't actually re-run anything." Comment from Neeraj H(@terranicgerald): @francois_marais_nz  That's a subtle one, and kind of a nasty class of bug because nothing actually throws ... ... agent you were running when that happened, always interesting to see if this shows up more with certain setups. Comment from Neeraj H(@terranicgerald): @ella_reyes1 That experience got me generating a JSON dependency graph that tracks which edits cascade

... huge." The value here is frontloading the signal, you get behavioral data before the architecture document is even finished. CI pipelines as a first-class use case. Resource saturation and cost surprises caught before staging. Lambda cold-start cascades. Overprovisioned RDS instances. NAT Gateway billing accidents. These are exactly what the simulator surfaces, and they're exactly the things that don't show up until a bill arrives or a load test reveals them. Chaos engineering with resilience scoring. Inject ...

... actual insight, and it's bigger than it looks: making the notification say why it's waiting is what turns a ping into a triage. "Done" makes you tab back to find out what "done" means. "Needs approval vs finished vs

... adds [YOU]: / [OTHER]: speaker labels. Dual-Cmd forces a tip mid-session. Notch escalates to always-on-top with a 3-second watchdog. New Knowledge Graph overhaul: Floating detail panel, click-through entity relationships, in-app entity deletion (cascades to triples), liquid-glass zoom controls, flicker-free node drag. New What's new modal: First launch after auto-update opens release highlights with a hero matching the website's auth screen and a link to the full changelog. New Welcome ... ... side-panel pattern we already use for file previews so it doesn't add a new UI primitive. Probably the very next thing on the roadmap. Two quick questions if you've got a sec: 1. Per-message receipt vs

... intuition is what pushed me to build Cascode, alone, around my full time job. It s a browser-based canvas where you assemble real AWS architectures, watch live message flows run through them, inject failures, and see what cascades. No setup, no cloud bills, just a live environment to build and break things on purpose. I m early and actively building. Real-time collaboration, advanced search, and a lot more are on the roadmap: cascode.uk/roadmap Try it free ...

... makes this even more urgent. More agents running in parallel means more third-party skills executing simultaneously, which means a larger attack surface compounding faster. One compromised skill in a multi-agent workflow doesn't just affect itself, it cascades. "In AI we trust" is the right question. The answer should be: trust, but verify. And right now almost nobody is verifying. Comment from J.D. Salbego(@jdsalbego): @madelinelawren   @fmerian  100%. With the abilities AI agents give humans ...

... fulfill my requirements. I will not completely rely on AI for my very important papers and research in my work. Comment from Briana Pole(@briana_pole): For me it's less about the data category and more about authority vs. visibility . I'll let AI know a lot calendar, to-dos, logistics but I don't want it acting irreversibly without me confirming, especially with money or health stuff. Honestly, what builds trust for me isn't the sensitivity ... ... taught me this the hard way: an AI that gives a confident-sounding answer when it's actually guessing is more dangerous than one that says it isn't sure yet. We ended up building explicit confidence grading (Verified vs

... Most AI systems assume time doesn t matter. But in production: delays compound, timeouts cascade, and retries change behavior. Nothing is wrong in isolation. Everything is wrong together. Time is a first-class input, whether we model it or not. Curious to hear from this crowd: Where has timing caused unexpected behavior in your system? Comment from Bhavin Sheth(@allinonetools_net): So true most of my AI bugs weren t logic errors, they were timing issues stacking up silently until ...

... enable "hyper-contextualized codebase understanding, grounded in precise code navigation". Or put another way: a map to help find your way in a thicket of vibed spaghetti code. I especially appreciate that you can reference a codemap directly in Cascade, Windsurf's coding assistant: What do you think? I'm pretty sure other tools like this exist but it's nice to have this integrated directly into an ADE. Comment from Sanskar Yadav(@sanskarix): Codemaps brings understanding front and center ...

... state through versioned memory, so forking or chaining doesn t break flow or history. Load management: Lock-free scheduling and atomic counters in Rust handle bursts without queuing collapse. Graceful recovery: Circuit breakers + classified retries isolate faults before they cascade. Would love to dive deeper, your work at Theysaid sounds like it overlaps a lot with what we re solving. How about a quick chat sometime this week to compare notes? Comment from Sanskar Yadav(@sanskarix): Absolutely! Making AI reliable ...

... complex code. It ll probably take a few months to hit the kind of polish we saw with Replit s Agent3 setup more recently. In the meantime, it s worth double-checking your database replication while monitoring the Cascade auto-scanning and code functionality as it may fail solid sandboxing before pushing live. Comment from Sanskar Yadav(@sanskarix): Looks like it might still be rolling out gradually? Sometimes these new features take a few days to hit all accounts. Worth ...

... www.youtube.com/watch?v=-7gm8mST9QU Windsurf Wave 12 Released! This is a major refresh that brings the first integrations of Devin's intelligence and capabilities directly into your Windsurf IDE: New UI Design : Fresh Chat and Cascade panels make it easier to scan agent actions and understand tool calls DeepWiki Integration : Hover over code symbols for AI-powered explanations (not just basic type info). CMD/Ctrl+Shift+Click to open detailed explanations in the side panel and add to Cascade

... Cascade has gotten better, we noticed that the tasks users are giving it have gotten more complex. Instead of just asking Cascade

... Claude 3.5 Sonnet levels of tool-call reasoning while being cheaper to serve. It will be available to all paid users for a promotional period of 0 credits per user prompt. SWE-1-lite : A smaller model that replaces Cascade Base at better quality. It is available for unlimited use to all users, free or paid. SWE-1-mini : A small, extremely fast model that powers the Windsurf Tab passive experience for all users, free or paid. These custom models ... ... cheaper for Windsurf users than the frontier labs companies offerings and replace Cascade

... credits a month, a user will get 25. With GPT-4.1 and o4-mini still on a discounted 0.25 credit rate, this corresponds to 100 such prompts every month. Unlimited full agent : We are also going to give unlimited Cascade Base model usage in Write mode to our Free users. Previously, unlimited Cascade

... mentioned tools for over a month, and are now at a stage to decide to purchase a Pro account from any one of the tools. We prefer Codeium right at this moment given it's likeliness to VSCode, the cascade feature in it's awesome. Have any one of you taken the Pro accounts from anyone of these? Please suggest us a better tool by explaining it's Pros and Cons. Looking forward to the helpful responses. Thanks in advance ...

... doctrine. In the AI era, it is increasingly hard to know where IP theft begins and ends. What happens when a creator uses an LLM that was trained on pirated works unbeknownst to them... Consider the cascade of complex questions this raises: Should the LLM maker be liable? The creator? Both? How do we establish the chain of causation in AI-assisted creation? How would courts measure financial harm from derivative AI works? What constitutes "substantial ...

... some improvement in the future. Comment from Rajiv Ayyangar(@rajiv_ayyangar): @kwindla   @matistski  while a true speech-speech model sounds like it opens up many possibilities, from an end user standpoint I m personally bullish on the cascade model. I ve said this before but I think voice-in/text-out is the highest bandwidth voice interface, currently only explored with transcription apps. I d love to see more of these in the wild. We read way faster than ... ... capital (although capital isn't the main bottleneck!) on bringing amazing researchers/research engineers (for any researchers reading - if you are interested - message me!) and compute. (2) We think you can likely bring more of that emotional context in that cascaded

... been using @Spacemacs for a few years now. Recently, I started using @VS Code to "see what I was missing". I liked that linters and such seemed to work a bit smoother out of the box. I kept hearing: VSCode is dead, Cursor is the way. Then I started hearing: Windsurf is the way, Cursor is dead. Today I used Windsurf for the full day. Some raw thoughts: I like the icons! Treating it like an "op autocomplete" was pretty ... ... would have been nice if I could say "chill out on the autocompletes until i finish this thought", maybe like a toggle somewhere? does this exist already? the center pop up menu is cool. how do i make the cascade

... editor journey has been: 1. Tried out Github copilot in VS Code. Loved it. 2. Then tried out @Cursor after @bernatfortet kept talking about it. Immediately switch to Cursor. 3. Then tried out @Windsurf after @lagap discovered it. Immediately fell in love and switched. Anecdotally, everyone is talking about Cursor. And no one is really talking about Windsurf. On Product Hunt, Cursor has 845 Shoutouts. Windsurf has two? In my experience, Windsurf is the leader at the moment with ... ... fine doing the steps my self. The real differentiator isnt really the IDE anymore. its the model. And I use o1 api key. It costs me $20-$50 a day. But its like working with a senior dev vs