Mustafa Najoom

Join us at TradingTech Summit. If you're attending and would like to meet up, send us a DM. #networking #tradingtechsummit

@jaredsuniverse true

Distribution = Moat Attention = Distribution Therefore: Attention = Moat Been true for a while. Market is just catching up.

AI cost < one week of a freelancer. Output = 20 extra hrs of research + analysis every week. Most founders are doing the wrong math.

@samwoods Team capacity × AI = A different company Most people are still pricing it like software, not compounding it like leverage.

I asked an ecommerce founder last week what would change about his business if his team had an extra 20 hours a week of research and analysis capacity. He described a completely different company. The decisions would be better, the launches would be faster, and the positioning would be tighter. Then I showed him that the AI setup to get those 20 hours costs less than one freelancer for one week. He'd been thinking about AI as a line item when the better way to think about it is as a capacity multiplier on the team he already has.

@APompliano Smarter at specific tasks inside defined contexts, yes. But they still need a human to decide what problem is worth solving in the first place. That gap is closing though, and you're right that most people haven't sat with what that actually means.

The uncomfortable truth is that AI agents are smarter than humans. The ramifications are wildly under-appreciated so far.

@plainionist The compiler as the world's most patient code reviewer that never gets tired of explaining why your memory management is wrong. Actually makes sense.

Hot take: Rust might be the best language for vibe-coding. The compiler provides strong feedback and guardrails. And AI thrives on guardrails.

@FedericoNoemie The 1 or 8 rule is funny but it holds. Two founders sounds balanced until someone has to make the call at 2am and both of you are equally right about opposite things

ran into a founder who had a co-founder and is now solo. he jokingly said startups should either have 1 decision maker or 8, anything else is a bad middle ground. every great startup becomes 'solo founded' eventually. you think jobs, not wozniak. zuck, not moskovitz. Airbnb IS brian chesky.

@geoffreywoo If Claude can navigate any UI without an API, your competitor is just as exposed as you are. The founders who win will be the ones who build on top of that capability before they get replaced by it. Distribution, trust, and proprietary data are still moats, not UI complexity.

psa to founders who think anthropic's computer use is just another feature: it's literally the death certificate for your entire saas category when claude can navigate your competitor's ui without apis, why would customers pay for your "streamlined workflow" your moat just became a puddle

@CasJam Two steps before you even pick a problem: 1⃣Spend a week just breaking things and asking dumb questions, you need to know what's possible before you can see which problems are worth solving. 2⃣Then find the real problem. The project teaches you everything after that.

"Where do I start with AI if I have zero coding experience?" Honest answer: forget the tools for now. Start with a real problem you need solved. Let AI guide you through building the solution. Your first project teaches you more than any tutorial ever will.

Distribution ate the product. The taxi wasn't bad because of the car, it was bad because of everything around the car. Uber fixed the wrapper not the ride.

@asaio87 The "too dangerous to release" line has become the best marketing in tech. Ship it anyway, call it controlled access, watch the waitlist grow.

For the past two years we have the same AI Each update was just incremental. It’s like they change the color, the box, have it a bit better and then sell it back to you like it’s something so good it’s dangerous to even release it. Sheep get excited. Only to find out that the next one will be the real deal. And then the next one And then the next one again

@kylegawley Fair shot. Though 'too dangerous for the public' and 'accidentally leaked our own code' being in the same announcement does require some mental gymnastics to sit with.

Everyone is buying some hype BS about a new model being able to uncover any security issue and it’s too dangerous for the public From the same company that leaked their own source code last week

@Lukealexxander The people who started in January don't just know more tools, they've built intuition for what works and what doesn't. That compounds in a way you can't close by watching a YouTube tutorial in month 7.

If you thought AI was moving fast from January until now you’ve got no idea It’s going even faster than I had guessed at the start of 2025 You’ve got probably 6 months to catch up before a real gap starts being created from those using it vs not

@businessbarista Nothing says 'we're serious about AI' like forming a committee to decide who owns the committee😅.

The person who owns AI for your company says a ton about it. CEO: "This is a inflection point in our company's history. We need to do this or we'll fall behind" CFO: "We need to squeeze $50m in EBITDA and this is how we'll do it." CTO/CIO: "AI transformation is a technology problem & we need it to be owned by our technology leader." CISO: "This is a data security nightmare, we need to make sure the inmates don't run the asylum." CHRO: "This is a people/process issue, AI is just small part of what is wholesale organizational change." Head of AI: "Former CTO or CIO with a fancy schmancy title."

@TheGeorgePu Feels less like exclusion and more like controlled rollout. With something this powerful, limiting access early is about safety and iteration, not just privilege. But yeah, let's see whether access broadens or stays concentrated, that’s what will define the impact.

Anthropic just announced their most powerful AI model ever. They also announced you can't use it. Only select companies get access. The richest companies get the smartest tools. Everyone else waits. This used to be how oil worked. Then land. Then capital. Now it's intelligence.

@trikcode Wild. This isn’t just speed anymore, it’s catching what humans missed for decades.

So Anthropic built a model that found a 27-year-old bug in OpenBSD. Twenty. Seven. Years. Millions of developers looked at that code. Millions of test runs. Thousands of security audits. Claude found it overnight. For $50. And Anthropic said "this is too dangerous to release."

@ALEngineered The asymmetry is the real problem. Offense only needs to be right once. Defense needs to be right every time. And most of the people who need protection aren't in the coalition.

You guys realize that Claude Mythos can’t be ethically released to the general public ever, right? That is, we just have to wait until the entire internet has been patched of all critical exploits, and all future code is forever scanned going forward. So no software should be released until it has been scanned by Mythos. But you have to be part of the handful of companies that have access to it. We are in a genie-out-of-the-bottle moment. When there’s a new major 0-day exploit, teams of agents will race to compromise systems while the means to stop them will be dependent on whether you are in the club or not (you are likely not in the club)

@toddsaunders Stage 5 is the funniest part because their family thinks something is wrong with them.

The funniest pattern with blue collar builders. Stage 1: "AI is for tech people, not for me." Stage 2: A friend shows them something they built. They're curious but won't admit it. Stage 3: They try Claude Code for one small thing. A unit converter or a simple form. Something they'd be embarrassed to tell anyone about. Stage 4: It works.... They stare at their screen like they just pulled a rabbit out of a hat. Stage 5: They disappear for two weeks, staring at their computer screen more than ever. Stage 6: They resurface with three monitors, a custom domain, and a full product demo for software that doesn't exist anywhere else on earth, because they understand the long tail nuance of the problem. I've watched this happen with plumbers, foresters, roofers, HVAC techs. The trade doesn't matter, and the arc is identical. The gap between "I can't build software" and "I just built software" isn't skill anymore.... It's one working demo. One moment where the thing you described in plain English actually appears on your screen and works. After that, there's no going back. It's funny because Silicon Valley spends millions on developer evangelism. But it turns out the best evangelism strategy is a contractor showing another contractor what he built over the weekend. Let the builders build!

@tomfgoodwin The market fragmented faster than anyone expected. Different tools for different workflows now, not one default anymore.

Pretty sure everyone I know stopped using ChatGPT a while ago Everyone uses Gemini or Claude or Grok now. It changed fast. ChatGPT is for the olds.

@DanielMiessler A model chaining vulns it was never trained for isn't a cyber story. It's a general capability story.

We’re missing a much bigger point on Mythos. It wasn’t even trained specifically for cybersecurity. It’s just that much better at doing work in general. It’s that good at cyber because it’s that good at everything. What do you think this is going to do to knowledge work? Mythos can chain multiple low and medium vulns together to create a high or critical. This is a task that far less than 1% of cybersecurity experts have ever done. Hell, probably less than 1% of all pentesters. So if it can do that, how do you think it’ll do at sending emails, doing analysis, writing reports, and the other 99% of everyday knowledge work? Do you really still think that Chris from Idaho has any chance competing against AI for a knowledge work job? In six months or a year, there will be very inexpensive models that can do knowledge work almost as good as Mythos. So companies have the choice of paying Chris $84,000 plus a whole bunch of benefits for 40 hours of mediocre work, or they can pay probably $100-$1000 for an AI that can do 10-1000 times the work per hour and that works 24/7. This Mythos announcement is getting attention because of cyber, but the real story is work in general.