since you guys loved the exploding tesla..
I used GPT-6 Astra to create a 3D website that pulls apart the male anatomy into 2,234 modeled pieces!
we are in a renaissance of learning
My two cents on this is:
If you’re feeling left out and don’t know what direction to take:
https://t.co/dikgX6rOyc’re not alone. The majority of the noise is skewed toward engineers who either directly work on these agents or have an investment in them, so they can spend more of their time exploring them.
2.If you’re unsure where to start, my advice is to go back to the very foundations. Forget whatever you hear or see, ignore the noise, log off Twitter if you must. Go build something that exposes you to the fundamentals. Read, implement, fail, repeat. The foundations of software engineering have never been more important than they are now. That’s how you distinguish yourself.
But Temi, I’d be left behind. No, not really. At least, I don’t think so. Here’s why:
•We’re just getting started, and today’s “hot cake” will be replaced in 3–6 months. New ways of working with agents will emerge. This year alone we’ve had MCP, skills, A2A protocols, etc. It’s obvious the field is still shaping up, and nothing is quite concrete yet.
•With solid foundations, you’d be better equipped to understand what is happening or what the agent is doing. You’d be in a better position to decide what requires you in the loop and what can be YOLO’d by the agent. You’d also be able to improve and grow at an accelerated rate.
3.Leverage the agent as your senior engineer while learning. That’s an unbiased engineer with billions of contexts on engineering. Ask it to review your work. Ask it to show you different approaches to doing things.
4.Most companies are still playing catch-up for good reasons. The tech isn’t there yet to fully give it autonomy. We’re still not close to self-evolving agents, and the justification of the cost vis-à-vis the results isn’t there yet. What does this mean?
There’s still a place for engineers with strong fundamentals, and companies are still looking for them.
People who can reason about systems, make trade-offs, debug when things inevitably break, and take ownership end-to-end. Agents can help you move faster, but they don’t replace judgment. Someone still needs to define the problem, put the right constraints in place, and be accountable for the outcome.
Until the tech is good enough to reason about context, evolving requirements, and long-term system health on its own, there’s a real place for engineers who understand what they’re building and why. And honestly, that’s where the leverage is.
So ignore the noise, build more, get your hands dirty.
@_iamtpo this advice is fine for students/juniors.
but imagine telling this (‘go back to foundations’) to the likes of karpathy. there really has been a huge shift and seniors/top talent do need to adopt their working models.
The #VeryDarkMan hack is a cautionary tale in balancing transparency, security, & privacy. 🚨 160M out of 180M was stolen due to potential critical flaws in app permissions, Banking APIs, & data privacy. Here’s the breakdown of what MIGHT have gone wrong 🧵⬇️
- This is all under the assumption VDM is not catching cruise or actually take the money himself:
🔑 Lessons Learned:
1️⃣ Limit app permissions: Access only the data needed. 📱
2️⃣ Restrict APIs to read-only for tracking funds. 🔒 3️⃣ Publish anonymized data to protect privacy. 🛡️ 4️⃣ Vet developers carefully & conduct security audits. 🔍
5️⃣ Secure sensitive devices with strong protection. 🔐
Transparency, security, & privacy can coexist—but only if designed properly. 💡 The #VeryDarkMan hack is a wake-up call for anyone handling sensitive funds or public trust. Let’s do better. 🌐
I hope we now see how this hack might have happened!
The #VeryDarkMan hack is a cautionary tale in balancing transparency, security, & privacy. 🚨 160M out of 180M was stolen due to potential critical flaws in app permissions, Banking APIs, & data privacy. Here’s the breakdown of what MIGHT have gone wrong 🧵⬇️
- This is all under the assumption VDM is not catching cruise or actually take the money himself:
While well-intentioned, this system introduced vulnerabilities:
1️⃣ Over-Permissioned SMS Apps:
Access to ALL SMS messages, not just transactions.
Enabled reading of 2FA codes, reset links, & more. 🔓
Exposed sensitive info outside of banking alerts.
2️⃣ Misconfigured Banking APIs:
A quick look at Zenith's APIs (publicly available) also shows it can perform transfers, and not just view transactions.
- API keys might’ve been hardcoded or leaked, creating easy targets. Might have been used to make unauthorized transfers! 💸
3️⃣ Privacy Breach: The app published sensitive depositor details:
- First & last names. 👥
- Transaction amounts. 💰
This violated privacy laws & exposed donors to: Phishing. 🎣Identity theft. 🕵️♀️Scams.
Transparency ≠ Publishing personal data. 🔐
4️⃣ Insider Threats (very likely): The developer’s role raises questions:
Malicious intent? 😈 Backdoors or intentional exploits.
Negligence? 🧑💻 Poor coding left the system open to attacks.
5️⃣ Phone Security Risks: Someone got physical access to VDM's phone, installed a malware (very unlikely IMO):
Malware/phishing could’ve granted remote access. 🛠️
Banking tokens, SMS messages, & personal data became low-hanging fruit. 🎯
It's crazy how folks can just conclude without even seeking any further context. Even with the current context, is this the conclusion you can come up with?