Dima

[Dima @0xpandev]

/trade

[cyber•Fund @cyberfund]

https://t.co/lLuW5EicJh

[Dima @0xpandev]

/watch

[cyber•Fund @cyberfund]

Building AI-Native Startups https://t.co/magiBXxkRK

[Dima @0xpandev]

/join

[cyber•Fund @cyberfund]

There are thousands of threads on building startups with AI. But what if you could watch someone actually do it?

On June 1, four operators share their screens and walk you through the AI-native workflows running their companies right now.

Real systems, live demos, no slides.

[Dima @0xpandev]

@Lomashuk /apply

[Dima @0xpandev]

just /apply

[cyber•Fund @cyberfund]

500+ applications in the first 24 hours.

Solo or team.

/apply

[choly 求理 (Ж) @zeroxcholy]

ai native founders, small team, conquering the world

[Dima @0xpandev]

Do the impossible.

[cyber•Fund @cyberfund]

Introducing the Monastery for AI-native founders.

A single builder can now outperform a publicly traded company.

$2 million. 12 weeks. Do the impossible.

[Dima @0xpandev]

🛕

[cyber•Fund @cyberfund]

Loading...

[Dima @0xpandev]

The barrier to building useful things has never been lower and that's genuinely exciting. But the security model for this new OS is still being written

[Dima @0xpandev]

Been thinking about LLMs as a new type of operating system. Instead of binary commands, they accept words as input, and skills and workflows become programs written in plain text

[Dima @0xpandev]

But the most interesting part is the context layer. This OS reads everything you feed it, and that's also its biggest vulnerability, a malicious string inside a document or a tweet can hijack the agent mid-task

[Dima @0xpandev]

@litocoen @ethena Congrats!

[Dima @0xpandev]

One I've been exploring: garak by @nvidia

Open-source LLM vulnerability scanner. Tests for prompt injection, data leakage, jailbreaks, encoding attacks.

Basically a penetration test for your AI setup. Run it before someone else does.

https://t.co/p8NnKwWHCN

[Dima @0xpandev]

Everyone's rushing to plug AI agents into their workflows. Very few are stress-testing what those agents will leak under pressure.

Security of AI setups is becoming a hot topic and for a good reason

[Dima @0xpandev]

Thinking about security at the design stage is crucial. Separate sub-agents by context. Don't let your GitHub agent touch your Telegram. Don't let your content agent see deal terms.

But even with solid architecture, there are tools that can help harden the system further.

[Dima @0xpandev]

Stablecoins at checkout are shifting from a cool option to something consumers just expect. Like card terminals everywhere.

One of the things I'm genuinely most excited about in crypto right now.

[Dima @0xpandev]

CLARITY Act yield compromise just dropped. The last blocker for US stablecoin legislation is done.

Senate Banking markup expected this month. Over 100 industry groups demanding action.

[Dima @0xpandev]

The wild part is adoption isn't waiting for the law to pass:

Visa already runs 130+ stablecoin card programs in 40 countries. $1.5B/month going through stablecoin cards. Projected 3% of all US dollar payments this year.