Reuben

@SharpCoder what would be the correct approach if you're using Supabase with no RLS?

I love Chris and I truly respect his journey as an entrepreneur, but the real takeaway here should be DON'T USE RLS! There is no way on earth someone who's worked with RLS once in there life can recommend using it. I honestly thought it was common knowledge nowadays that RLS was a very bad pattern, and that no serious application should be relying on it.

@raroque happy to test this out

@ReubenKib it actually doesnt even need your repo, i just need a URL + credentials for a test account 👀

i have this personal tool i built to help me audit the security of my own apps 👀👀👀 its actually getting reeeeally good and i think i found a way to *safely* release it without people abusing it anyone want to help me test?

@migdoesit css key frames

@ReubenKib motion?

gemini 3.1 can cook

@internetvin When you premiere this, you should serve shawarma instead of popcorn

but I need to understand much more than I do.

the hype about openclaw needs to be studied

@evilrabbit_ I'll pay top dollar for this

▲

with codex spark you can build an app in 3 seconds

found the issue, cursor was making preview deployments automatically. lesson learned. loving @convex so far.

@rjgcoach @convex confirmed it's not for every commit. something else is happening, which is reassuring

@ReubenKib @convex Sounds like we should start worrying!

making the switch to @convex and haven't been able to find the answer in docs. I'm already at 30/120 deployments (every time I commit to GitHub there's a deployment). I must be missing something because I will be hitting 120 deployments soon.

can't wait for this to be available to everyone. long overdue!

@mikeysee @GeminiApp @convex Are the guidelines the ones you guys have in the docs for the coding assistant?

Okay the new @GeminiApp 3.1 Pro added to the @convex leaderboard and wow, straight to the top with a near-perfect score! The super strange thing tho is how bad it does WITHOUT the guidelines. I'm digging deeper now but I think this means that Gemini is REALLY good at following instructions but innately doesn't know all that much about Convex. So make sure you add the Convex guidelines or point to the docs if you are going to use this model!

Google dropped a “.1” update: Gemini 3.1 Pro (Preview). Point-ones are still marketing moments. But they’re starting to be more than just a campaign. For design work: better reasoning, token efficiency, and more grounded output = less drift across iterations.

@DavidOndrej1 second time a major outage in the past 3 days. not looking good.

OpenRouter is down. It's over.

it's happening again. both times I do user interviews, and @OpenRouter has a major outage.

Contra just announced a new payments product where you can publish products and sell to AI agents. That’s a clear signal of the next shift. Up until now, we’ve mostly optimized for one thing: human conversion. Copy, design, trust cues, testimonials, pricing psychology. But now we also have to optimize the funnel for agents. And agents aren’t “convinced by clarity” the way humans are. They buy based on the foundation of the funnel + product: - Can they complete it through tool calls? - Is the offer represented in something they can control (structured text/files)? - Can they validate what they’re getting without a meeting? The playbook is still evolving, but the direction is obvious. Next time you build your funnel, test a few agent-specific implementations and see what actually converts best. This feels like the early days of human conversion all over again. Just with agents.

AI is making building more accessible than ever. It’s never been easier to go from idea to app. The catch: a lot of new products are starting to share the same look. Same components. Same gradients. Same “AI startup template” vibe. When everyone can build, design becomes the moat. That’s why I’m building DesignFast. I want teams to ship fast without UIs that look like they came from the same prompt. In the AI era, design isn’t decoration. It’s how users understand the product. And whether they trust it.