... expect (a web OS is meant to be used from a desktop browser anyway), although we did make the web OS responsive and optimised for mobile it still wasnt enough, the huge variety of mobile devices handle networking, live sockets, background sessions very differently from each other, which we didnot expect, and it broke things we didn't even imaging could break. Since then we have shipped 60+ commits to fix all of those issues and still continue to ship ...

... waits for the endpoint to name the remedy in its own error , and retries exactly once with the value it asked for tools still attached, because answering without the search results would rather defeat the point. Housekeeping Two scripted-socket tests for the retry path, which previously had a proud total of zero: one pins that the second request really goes out carrying the same tools, the other that an endpoint refusing even under none gives up instead of hammering ...

... same rotating Microsoft refresh token at the same time and both lose. Token requests got a timeout and a reused HTTPS client. Previously a black-holed auth endpoint could hang around for as long as Windows kept the socket alive. Now it gets 15 seconds and then it's out. Discovering a dead connection dropped from ~22 40s to ~4 6s. One tiny NOOP up front instead of twenty seconds of staring at "Loading ". Bold, we know. Fixed Ordinary mail ... ... where prefetching risks writing files to disk for a message nobody opens. What I can't hand you is a benchmark across real-world connections. The one hard number I have is the adjacent fix: discovering a dead fetch socket

... years? 4 This is where I want to bring concepts from software and hardware engineering into the clothing industry LTS and APIs. What if we freeze the standard and support it? We already have things like SATA or CPU sockets. Why not build a similar concept for clothing? Then the API at the point of contact with the customer is divided into a physically immutable part (color, material, design) and a flexible part (logistics, ingredients). Let's start with ...

... space when I am in a new place and want to tap into the community there and join events. Cafes can be nice but you re battling the elements, rowdiness, chatter, background music, uncomfortable chairs, and a lack of sockets. Plus, if you re alone, going to the lavatory is an issue in terms of safety. Home always wins, but that s just ...

... rather than hold the line forever. A stalled job that took the cautious path is recoverable. A stalled job that just sat there is a support ticket. Comment from Dale Mooney(@dalemooney): @alex_amancio_candoa_   @aadilghani  Constant versus invariant is the useful thing here and I am taking it. One addition: you cannot separate them by observation, but you can by deliberately violating the constant once. A property that survives a test faking an unattended caller ... ... which is bounded, and cannot fire when nobody started anything. Your last seen sweeper is already the mechanism. It is sweeping into the wrong bucket rather than lacking the signal. Comment from Dale Mooney(@dalemooney): @aadilghani   The gate versus

... would be best, definitely. From my experience, the most reliable way would be to ensure the search and installation tooling won't allow AI to discover and install fake\malicious packages. I think this product is doing exactly that: @Socket - there's a Claude plugin, adding a requirement to check all new packages against their scoring should help. Comment from Praveen(@praveen62): @artk  I completely agree that tools like Snyk and Aikido are valuable as part of the defense ... ... introduces an unexpected package, I'd like to know which agent session suggested it, why it entered the workflow, and whether it passed basic trust checks before installation. I see registry scanners and provenance as complementary layers rather than alternatives

... doorbell. It rings the same whether it's a package or a fire. That's basically the gap Pushary fills, same hook idea you've already wired, but the event carries a real handoff (done vs blocked vs

... been away. Faster, by a lot Spawning a Claude agent on macOS used to stall for ~23 seconds when more than one came up at once. We moved the PTY off Tauri's IPC channel onto a loopback socket, so now ten agents start near-instantly. Windows got its own overhaul in 0.25.0. The "Not Responding" freezes when you opened several projects or switched branches are gone, because every git and filesystem call now runs off the UI thread. Large ... ... batch your agent runs and dread spinning up a new one. Near-instant means you just do it, ten times a day, without thinking. The fix was moving the PTY off Tauri's IPC onto a loopback socket

... Krishna Kant Singh(@kkxingh): @rohanrecommends   Background tabs are kept alive but idle the terminal rendering loop suspends when a tab loses focus, so there's zero GPU/paint overhead. Each tab holds: 1 persistent SSH connection a lightweight open socket, near-zero CPU when idle 1 terminal buffer ~2-5 MB depending on scrollback history So 10 open sessions 10 sockets

... commonly used libraries are. I mean, web3.py went 5 versions with basically a fake websocket client. Feel free to dig up the voluminous discussions on github about workarounds to this, except now that it's no longer a fake socket, all that easy MEV you could have made that was basically free money for years pretty much no longer exists for you to get. It's all use-case dependent, but I rarely find copilot or iwhatever AI du jour ...

... which stressed me out a bit. Although I realized I never hit the threshold where it started charging me more than my normal subscription. So I think it was just a psychological thing. It had repeated errors around web socket connections and dealing with notifications on iOS. I had to get deeper and deeper into these myself, and in one case, teach it how to do notifications properly on iOS. It turned out that Replit's information was outdated ... ... next week or for Agent that should see significant improvements by default, but some additional context that might be helpful to you: 1. Why it happens: The Agent (language models generally) is worse at fixing something it broke vs