Marcel

@thsottiaux Is @AmpCode also affected by this or does it have safeguards in place?

On file deletions. We’ve investigated a handful of reports where GPT-5.6 unexpectedly deleted files. What we have found is that this most commonly occurs when: - Full access mode is enabled and codex is run without sandboxing protections, including without auto review being enabled - The model attempts to override the $HOME env var to define a temporary directory. - The model makes an honest mistake and mistakenly deletes $HOME instead. This is of course not how we want the system to behave, even when a user operates the model in full-access mode without the safeguards of our sandbox or without using auto review which checks for these kinds of high risk actions and rejects them. We are taking steps to mitigate this risk including by updating the developer message, guiding more users towards safer permission modes, and adding additional harness safeguards. Even though this happens extremely rarely, we’ll share a detailed post-mortem in the coming days that goes into more details and what we are doing to minimize risks further.

@stolinski @thdxr @swyx @sytses Makes you wonder why they didn't also invest in some fake comments...

@thdxr @swyx @sytses This is always the biggest tell. 11 comments with that many views is egregious

as someone who does a lot of dev community and dev youtube this pace of growth has been one of the greatest mysteries to me because clearly there's something to learn here @sytses i'd love to talk to whoever was the growth guy/gal at Kilo!!

@melvynx nice! what tool is this?

Built the same timezone checker app with GPT 5.6 Luna XHigh VS Terra High VS Sol Medium Luna XHigh: → 1,061,469 tokens → 10 minutes → $0.37 Terra High: → 1,540,618 tokens → 8 minutes → $0.86 Sol Medium: → 1,642,413 tokens → 10 minutes → $1.80

@MaxRovensky right def looks nicer, colors in left are way off in the foreground

okay would you mind telling me which of these photos is Kodak Vision3 500T and which one is iPhone 17 Pro? cause what you're describing is a skill issue

@jrysana what happended to this idea...?

The fact that there’s 30 options is kind of insane

@YashSolanki_ @ibocodes you can use logpush to newrelic (generous free tier) let codex with computer use create a custom dashboard for you

@ibocodes What about monitoring/analytics. That requires seperate paid plan on cloudflare. Do you have anything else for that to use?

my entire 2026 stack runs on cloudflare and I pay $5/month for it - workers (compute) - d1 (sql db) + better auth - kv (cache) - r2 (object storage) - queues (background jobs) - hyperdrive (db acceleration) - email routing (sending) - + waf & ddos built-in

@TheKoopaGuy Madeira, Portugal 🇵🇹climatico.app/en/2267827/539…

Is there anywhere on Earth that modtly stays between 15°C and 20°C and rarely goes above 25°C? That's where I'd spend my life if I could

@trevin just mention the skill name, no $ needed

Every time I go to invoke skills in Codex, I’m reminded that I want to strangle the person at OpenAI that decided it was a good idea to use a $ vs a /. This was not a thing to innovate on.

@TejasKumar_ Can you give examples of what they can do that currently available models cannot?

ive been using mythos-class models and they are genuinely agi i am not surprised at all the gov pulled them. wild.

@levelsio @sachasayan @Andercot can this fit rescue crew, patient on stretcher, medical equipment, winch, gear etc and still fly 4-5 hours like a helicopter?

@sachasayan @mrcldev @Andercot

When you first see a helicopter as a kid you're like "oh I get it, like a plane pointing up" And then later you learn it's just so much more retarded than you could've thought.

@sachasayan @levelsio @Andercot Are they used for emergency transport already? I think a huge problem is still payload and flight time

@mrcldev @levelsio @Andercot Hexacopters/octocopters.

@levelsio @Andercot What's the better alternative for emergencies and rescues then?

@Andercot Helicopters are a terrible idea

@RimasXYZ same

is openai down???

@piodubro @shadcn vite dev --host

Hey @shadcn, I've used your work across a bunch of projects in the past and it's saved me a ton of time. Not sure what your workflow looks like, but if Proxee could be useful to you, I'd love for you to try it and I'd be happy to give you and your team access as my way of saying thank you 🙏 Proxee (proxee.dev) is a macOS menu bar app that turns your phone into a live mobile preview for your local dev server, letting you monitor your site on a real device and catch responsive UI issues as you work. No code. No adapters. You set your localhost port, scan a QR code and pair your phone. That's it.

@salauddin_n @AmpCode You are complaining about getting free US$150 every month?

Been using @AmpCode for long time It's great tool and its harness also good but recently they reduced the fre quota 5 dollar which is bad. May they should introduce something basic plan like around 5 dollars where users can get come credits and use it instead pay as you go model

@CootJimbob @GeorgiaPics just look at google maps

@GeorgiaPics This is literally AI it does not even remotely look like this. People in the replies are retarded

Atlanta from the sky

@RhysSullivan if you don't want to trust random chrome extensions to block it, vibe code your own, its just a few lines of code

youtube has completely ruined their homepage with shortsslop

@CloudflareDev just letting you know, the german version of the workers website reads poorly (seems auto-translated). It might be better to either have it reviewed by a native speaker or simply keep the content in english, as your target audience will likely understand it

@nutlope Nice, thanks for sharing the code. I will try it but I think my M2 16gb isn't strong enough to run a vision model constantly without slowing down the machine too much...

@mrcldev Totally fair feedback! The code will be open source & you can run it with a local vision model too

I'm building an AI app that lets you get deep work done. You share your screen, start a focus session, & get deep work done with real accountability. Will launch & open source the code in 24 hours.