Hafez

You may find different posts on different topics here. Don’t worry, that’s what happens when a developer becomes wise… and an investor 😄 I believe we can all grow in this beautiful space with mutual respect and trust. Without them, it becomes a jungle. We’ll all be judged one day, so let’s not rush to judge each other. Whether you’re just passing by or genuinely interested, you’re welcome.

@suryanox7 Haha needed so we keep our jobs.

@HafedDm This should be title for some netflix series 😭

What’s the most important lesson you’ve learned from AI?

@readswithravi Well said. And never stop learning from others.

“We don't have to be smarter than the rest. We have to be more disciplined than the rest.” — Warren Buffett

Honestly i bounce between a few. ChatGPT for thinking and drafting, Claude and Antigravity for the actual code and planning/challenging approaches. Which is kind of the problem, no single tool wins, you end up switching by task. That’s the whole reason i’ve been exploring a concept around it: adaptiveaimate.com

SOFTWARE ENGINEERS ONLY Which AI coding tool do you use most?

@PeterDiamandis True. Believe and GOOO.

The cost of starting a company used to be $5 million. Then $500,000. Then $50,000. Now it is closer to $500. The only expensive thing left is your unwillingness/fear to begin.

@sickdotdev Smart question! But that logic says no one should sell pickaxes during a gold rush. Building Claude captures value from every $1M winner, not just one of them.

If making $1M is possible with Claude every month. Why would Claude's engineers work at Anthropic?

@Prathkum Vibe coding is the proof. More productivity helpers every day. Still no replacement for someone who actually understands the system.

The wisdom of markets is that AI makes developers more valuable. The madness of crowds is believing it makes developers unnecessary.

Which AI should I use today? ChatGPT? Claude? Gemini? Grok?… The tools are amazing. The CHOOSING is exhausting. What if you never had to pick — you just describe what you need, and the right tool/model gets routed in automatically? A concept I'm exploring, based on a metaphor from the human body and a little philosophy in the age of AI. adaptiveaimate.com #AI

@jameshutch1ngs @PeterDiamandis 😂 you still can

@PeterDiamandis 34 - does this mean I'm cooked?

Somewhere right now, a 19-year-old with no funding, no degree, and no connections is building something, where you least expect it, with AI that will outperform a Fortune 500 company's best effort.

@DrCronkenstein @RoundtableSpace ha, i got way too into this topic 😄 bit greedy when it comes to my own project’s code

@HafedDm @RoundtableSpace Thanks bot

THE WINNER OF THE ANTHROPIC HACKATHON JUST OPEN SOURCED HIS ENTIRE AI CODING SETUP FOR FREE. 183 AGENT SKILLS, 48 SUB-AGENTS AND 79 READY-MADE COMMANDS. He spent 10 months on it, won $15K in API credits, then released the whole stack under MIT license.

@russellbrunson Hard work pays off.

You’ll think nothing is working… then suddenly everything compounds at once. That’s success.

The long-running task focus is the right bet, that's exactly where a single model starts to strain, early steps are simple, later ones get hard. Would love to see it pick the right intensity per step automatically instead of running one mode the whole way. That'd make "always-approve" feel a lot safer too.

Composer 2.5 is now available inside Grok Build. Composer 2.5 is a fast, highly intelligent model that excels on long-running tasks and following complex instructions.

@ChrissGPT Agreed, computer use is huge, it finally lets the agent actually do the thing, not just talk about it. The next question imo: which model should be driving each action? The best one keeps changing task to task. Capability is one mile, knowing which model to trust is the next.

I’m so happy Claude and OpenAI never stopped grinding on computer use. Literally one of the most important aspects of automating work

@garrytan All that scaffolding is really just distrust written as code. The moment an outcome is measurable, the cage isn't needed anymore. The future you're describing is less about less code and more about trusting intelligence where we finally can.

We're still writing code like it's 2013: no latent space, no intelligence. When there is intelligence we gate the agents like workers at a Foxconn factory. The future of software is just-in-time and is 10x less code because of the markdown. And the agents will be free.

Great list. But honestly this is also the trap for most owners, 25 use cases, each needing a different tool or model, and no time to figure out which. The real solution isn't more AI options, it's knowing which one to point at each task. That's the actual bottleneck for a busy owner.

The biggest opportunity for small business owners right now: Supercharging your business with AI. Here are the top 25 examples: 1. Customer service: Agents that answer the same 10 questions 24/7, so you don't have to. 2. Follow-up: Automated emails that go out the second a lead goes cold. 3. Bookkeeping: Categorizes expenses, tracks cash flow, flags issues automatically. 4. Hiring: Job posts, onboarding docs, and training guides written efficiently and effectively. 5. Proposals: Quotes and contracts drafted faster than all your employees combined. 6. Marketing: One piece of content turned into 30 posts across every platform. 7. Scheduling: Handles appointment booking and reminders (without a front desk person). 8. Reviews: Monitors your reputation and drafts responses automatically. 9. Research: Competitor analysis and pricing intel pulled together in an afternoon. 10. SOPs: Build your entire operations manual just by talking through your process. 11. Invoicing: Follow-up that runs automatically until the invoice gets paid. 12. Social media: A month of content planned (and written) for your brand. 13. Ads: Copy that tests itself and keeps improving without you touching it. 14. Inventory: Demand forecasting so you stop over-ordering and bleeding cash. 15. Legal: Lease reviews, contractor agreements, and basic docs drafted with insane efficiency. 16. Upsells: Identifies which customers are most likely to buy again. 17. Pricing: Builds margin-based pricing models you should have had years ago. 18. Vendor management: Drafts supplier contracts and tracks renewal dates for you. 19. Training: Answers new hire questions instantly so you're not the bottleneck anymore. 20. Location research: Site analysis, foot traffic, and demographics on demand. 21. Cash flow: Spots waste in your subscriptions and expenses automatically. 22. Communications: Internal updates and team memos written (and sent) in 30 seconds. 23. Forecasting: Revenue projections based on your actual numbers, not estimates. 24. Inbox: An AI assistant that triages emails and flags what actually needs your attention. 25. Competitive edge: Using all of the above while your competition still hasn't started.

@pmitu More of a concept I’m exploring: AdaptiveAiMate, an AI that picks the right AI for you. You describe the task, it routes to the best model automatically. No more ‘which one should I use?’ every time. adaptiveaimate.com

Builders, pitch your product 👇 (a lot of smart VCs read my threads)

Agree on the friction part. But i'd push back on "one superapp owns it", the superapp still has to decide which engine handles each task. The choosing doesn't disappear, it just moves under the hood. Whoever does that routing best wins the agentic layer, not whoever has the nicest wrapper.

I think the AI superapps will soon own 90% of the agentic layer more and more people won't use hermes/openclaw etc because Claude Cowork/Codex will offer 90% of the functionality with 5% of the friction

@Tim_Denning I'd say unrealistic in size, but realistic in the next step. The goal can feel impossible from here, as long as tomorrow morning you know exactly what to do. The dream pulls, the small step moves you.

The goal you should pursue should feel completely unrealistic.

@yashhq_22 To be honest, money is part of it 😅 but the real pull is enjoying the work and making things easier. Both can be true at once.

be honest. are you building because you enjoy it or just for the money?

@rauchg @vercel What i love most: it lives in iMessage instead of asking people to download yet another app. Meeting users where they already are is half the product. Clean stack too. Bookmarking this one 🙏

Beautiful example of a full-stack agent on @vercel. Great learning material!