This is crazy advice. They recommend that their users nuke their apps without a way to restore (deploy) them again. The fact that they didn't restore deploy functionality means only one thing: they have no clue what the attack vector is, and they aren't confident that secrets won't be exposed again.

The era of AI cybersecurity has arrived!

And that's why AI won't destroy or absorb software.

You just need to whitelist ONLY Cloudflare's IPs on your server. No need to take your site down. It's 100% secure

Vibe coder though he could dunk on PHP devs

URGENT! If you are running a PHP app, read this 🛑🛑🛑 Cloudflare did their part, but your system is STILL vulnerable to direct IP address access attacks. You must either restrict your web server to only allow Cloudflare IPs or update PHP immediately

Omg, people! I was busy this morning and most of the day, and I’m getting back on X and seeing this 🤯 I’ll try to respond to every message... I have 100s of notifications. “Let’s connect” rocks!

I wonder if this is an engagement stunt organized by Marc and Tyler. It's too entertaining to be real

I believe that many AI startups are going to find out that possessing an excellent product does not automatically imply that they have a good business. At the moment, we are witnessing a number of startups launching products which would have been seen as technically impressive only two years ago. These include AI writing assistants, research tools, customer support agents and coding tools as well as document processors. A few of them are actually useful and a great many have impressive demonstrations. But there's the issue: if your product is just a prompt, an API call, and a good interface, then why not have someone build an improved version next weekend? What would take place then if no one else copied you, yet OpenAI, Anthropic, Google, or another model company included your key feature directly into their own platform? I don't think the solution is to add more features, since your competitors also have access to the same models, the same development tools, and are also able to release software quickly. What I find more interesting is the kind of thing you can create that becomes more difficult to replace the longer customers use it. Imagine a company that embeds itself deeply into a customer's operations, understands their workflows, connects to their key business systems, and eventually becomes where significant work gets done. In that case, replacing the product isn't just about creating an alternative interface; it would require replicating its integrations, migrating data, understanding the customer's processes, and persuading them that the disruption is worth it. This is a far more interesting business than another AI wrapper. Certainly, none of these advantages are permanent since integrations can be copied, customers can leave, and even well-established platforms can become outdated. Yet they give you something another team cannot reproduce just by having access to a better model. In my opinion, it's easier than ever to create a useful product, but harder than ever to build a company with a defensible position. Perhaps that's the mistake many founders are making in this AI boom; they're putting all their effort into creating something impressive enough to launch, without considering why anyone would still want to use it three years from now. Your greatest competitor isn't necessarily another startup; it could be the next version of the model that turns your product into just one of its features.

This is probably the best feature OpenAI delivered in this one-month challenge (until now)

One thing I'm glad AI destroyed is selling overpriced, couple of hours-long courses online ($349-$1200)

I should probably start warming up Insta and TikTok accounts for my products Have experience with these 2? I never did anything professionally there

I'm not sure why, but for small accounts, "let's connect" posts have really high reach. If you're new here or building your audience, and you write interesting things or build interesting apps, let's connect. I love creating things and turning ideas into real products. If you’re a: → Developer → Designer → Entrepreneur → Maker → Creator → Someone who loves building things Let’s connect! Excited to see what you’re working on 🚀

This is interesting. I can't wait to get my hands on the repo.

The fact that companies are refusing to employ junior developers on the grounds of AI could end up causing them a far greater problem. All people want senior engineers who are able to design systems, make good architectural decisions, identify bad assumptions, and take responsibility when things go wrong; but if no one is willing to employ beginners, then where are such engineers supposed to come from? For many years there had been a simple way of proceeding: you began by doing smaller tasks, made some mistakes, learned from more experienced colleagues, and then gradually took on more responsibility. Nowadays, AI is able to carry out a great deal of the work which we used to give to junior employees, and businesses are beginning to wonder why they should employ them at all. In my opinion, that's a rather narrow way of looking at the situation. AI could in fact be the best learning tool that junior developers have ever had; a person who previously used weeks to implement simple functionality is now able to create whole applications, experiment with various architectures, look at unfamiliar codebases, and receive immediate feedback. A motivated beginner could therefore gain a great deal of practical experience. But there's a drawback. More code doesn't necessarily mean you'll gain more knowledge. You could end up being very productive without acquiring much engineering judgment if you let an agent make all the decisions and never question its results. That is where I believe companies have an opportunity: rather than employing junior staff to write boilerplate code and correct minor bugs, they should give them real problems, provide them with experienced mentors, and supply them with AI tools. This way, the juniors should be allowed to build things, make their own decisions, investigate any failures, and explain the reasons for their solutions. Certainly, mentoring does involve a cost, and companies have a motive to focus on immediate productivity. However, if all companies conclude that it's not worthwhile to hire junior developers, then we're going to end up in a rather tricky situation a few years from now when everyone still wants experienced engineers. It is impossible to create a workforce made up of senior engineers unless people are willing to invest in their development. I think that, if we use it well, AI could speed up that development process rather than make it slower.

Building in public is a sham - just an echo chamber. Prove me wrong.

This was by far my most popular post on X. I'm not sure why this is so controversial to say. It's economics 101

I’m Mirza I love creating things and bringing ideas to life as real products. If you’re a: → Developer → Designer → Entrepreneur → Maker → Creator → Someone who loves building things Let’s connect! Excited to see what you’re working on 🚀