SadMouse

Looking for beta testers. $1,000+ in rewards. @t2000ai a bank account for AI agents. Save and earn yield. Borrow against savings. Buy crypto and gold. Send payments. Pay for APIs with Sui USDC gift cards, physical mail, AI models. No API keys. No accounts. CLI or Natural Language. discord.gg/qtVJR5eH

That's cheap if we compare with prices in Europe.

Miami Solana Hacker House 2021

A few days ago I saw a post that I still think about. It was along the lines of: 'If you use AI to fix all your messages, you lose your personal touch. Better to post in your own wrong style and sound like you, than to fix everything with AI and sound like everyone else.' For a while I was fixing most of my tweets. Now I'm thinking my broken English maybe isn't such a bad thing. It sets me apart. And after all — if you met me in real life, I'd sound like my broken English, not whatever AI does to make it sound perfect.

It's crazy how a simple cold can ruin your ability to do any reasonable amount of work. It's just a runny nose and watery eyes. That's it. But noooo… you will suffer and won't do anything of value for several days because of that.

@SuiCommunity Depends. If I ship something worth showing and getting feedback on, that alone would make the trip worth it. If I have nothing to show, it'll be hard to justify the expense. Unless I decide it's worth going to hunt for job opportunities.

Will you be there?

Basecamp is back! 🥳 But Singapore is so far away ... 😩

@matteodotsui You love BTC price even more.

I love my MacBook Pro

Used Claude to add a feature to one of my mobile apps. It sort of did it. Then I started testing the logic and found a few issues. It fixed them. I tested again and found several more problems. Now I've decided to completely change how it's implemented. That should simplify the logic and potentially make the feature even better. This experience yet again confirmed what I've realized several times over the last few months: * AI is really great for quickly prototyping things, but it won't think of all the edge cases. * Many of us will still have jobs, because someone always needs to be there to define business requirements, test things, and orchestrate our new minions. Development has changed forever and we need to adapt. But the work is still there. At least for now.

I have a feeling that Cloudflare's Bot challenges (interactive, managed, JS) aren't nearly as useful as they once were. Is that really the case, or am I imagining it?

Display V2 is finally launching with Sui v1.68! We always wanted Display to support richer queries and be able to fetch dynamic fields or map attributes. Finally, we got there! 🧵

A new account for devs just entered the chat. Say gm to @suidevelopers 👋 This is where we break down: • How Sui actually works under the hood • What's new in SDKs, tooling, and infra • Who's shipping, what they're building, and how And the occasional dev meme. Join us!

If I want to buy 1 SOL as a business, what's the simplest way to do that? Without some lengthy and tedious KYC checks etc. I should get not only 1 SOL, but also proper paperwork (invoice). Bank wire or card. Something that works in EU.

@matteodotsui hello

No account should be under 10k followers Say hello, I will boost you

@resend Added Resend webhook handling. OK. It was not really me. Codex did that. In one go. Your docs make it easy for agents to integrate your tools. 👍

What are you building this weekend?

Too bad I live in the woods and there are no people to host events for. ☹️

Been testing Codex for the past few days to compare it with Claude Code. Honest take: it's really good. The Plus plan also feels noticeably more generous with token limits than Claude's Pro plan — I hit walls less often. Code quality? Hard to call. Both handle everything I've thrown at them — though I'll admit my tasks aren't particularly complex. More testing ahead. This is just a first impression.

@trq212 Any hints on how to figure out what effort level to use? Like - for planning use high, for implementing use medium. Or maybe for patches medium, for new features high. Or something else? Any hints?

Our data shows that for the average user on Claude code, medium effort is the best balance of latency, intelligence & usage limits. If you feel medium effort doesn't work for you, you can switch to high effort in /model. You can also use ultrathink to think harder for one turn

We're continuing to see strong growth and excitement for Claude products. We’re so thankful for all your support. Here are some things we're doing on Claude Code to make sure things are running smoothly.

@oliveeyay 1-2 weeks?

So now that GPT-5.4 is out, when do we get Opus 4.7 and Gemini 3.2? 😅🤖🚢

Sending emails to users, I was surprised to discover how many people don't care that their inbox is full and emails are bouncing. I consider email availability critical — assumed everyone did. Turns out many do not.