Jazz • Product & Motion Designer #NeoBank

@YDaftary True, but identifying the real constraint is hard when everything feels broken. Most founders only see it clearly in hindsight.

WHAT BUSY FOUNDERS DO: > Solve whatever lands on their desk that morning > Treat every problem like it's equally urgent > Measure progress by how many tasks got done > Stay in execution mode 24/7 without questioning what they're executing on WHAT EFFECTIVE FOUNDERS DO: > Step back and identify the one constraint holding the business back > Say no to everything that doesn't directly address that constraint > Measure progress by whether the business actually moved forward > Treat clarity on what to work on as more valuable than the work itself The founders I know who break through are the ones who figured out which problem to work on.

@lawheroezV2 Plausibility is the dangerous part. It lowers the instinct to double-check, which is exactly where legal work can’t afford mistakes.

Plausible legal slop created by general purpose AI tools is still slop. And only a lawyer can tell the difference. Plausibility is one of the biggest psychological illusions created by AI. That’s why the combination of legal AI and legal expertise is so key.

@JamesonCamp Most “AI implementation” is just tooling, not business transformation. Owning the P&L is what makes the difference.

There are 10k people selling “ai implementation” But how many of them have actually ever in their lives built a successful business, know how to figure out rev ops, management, back office, workflows before And truly owned a P&L Maybe us and 2 other firms here

@ConnorAbene True, but figuring out which 3–5 actually matter is the hard part. Most teams only know in hindsight.

Every business I’ve worked with follows the same pattern. 3-5 variables actually move the needle. The rest is noise. If you’re tracking more than you’re acting on, you’re wasting time.

@MartinGTobias Good filter, but some of the best investors won’t pre-commit intros that early. It’s more about how they behave post-investment.

Whatever round you are raising ask every investor this question: Tell me the three next round investors that you know well who you will introduce me to if we hit our kpis for this round? If you get crickets. Move on.

@TheGeoMethod Pushing for a decision works, but timing still matters. Force it too early and you just accelerate a “no.”

Following up is fucking pathetic. “Just touching base.” …Sounds like borderline sexual assault I remember a guy early on who used to call people and say: “I’m just touching your base.” His English wasn’t great. But what he meant is exactly what every follow-up says: “I’m waiting until it’s convenient for you.” But NOTHING that requires selling EVER happens when it’s “convenient.” If it were… They’d already fix the problem. Sales exists because decisions are uncomfortable. They cost money. They require change. They force people to face reality. When you “check in” politely… You’re waiting for a moment that will never come. Instead, treat every interaction with a client as an opportunity to move the deal forward. Ask the question you know they’re hoping you won’t ask. “Is there something here that doesn’t make sense yet?” “Are we actually doing this, or should I move on?” “Is there any reason this shouldn’t move ahead?” Now the conversation has somewhere to go. You either uncover the real problem… Or you get the decision. Both are better than another pathetic follow-up.

@andrewchen Partly, but it’s also incentive-driven. When distribution rewards volume over quality, “good enough” becomes the dominant strategy.

what people call "ai slop" is just AI generated content that's not good enough yet

This flips the frame correctly. Payments follow behavior, not the other way around, whoever owns the interface to real-world actions captures the value.

@ClementDelangue @Pinterest @Airbnb @NotionHQ @cursor_ai @eoghan @intercom Makes sense at scale, but most companies underestimate the cost of maintaining these systems long term.

After @Pinterest @Airbnb @NotionHQ @cursor_ai, today it’s @eoghan @intercom publicly sharing that they’re finding it better, cheaper, faster to use and train open models themselves rather than use APIs for many tasks. And hundreds of other companies are doing the same without sharing. Ultimately, I believe the majority of AI workflows will be in-house based on open-source (vs API). It took much more time than we anticipated but it’s happening now!

@fivosaresti GTM engineers look great on paper, but most teams don’t have enough volume to justify the complexity.

Sales team setups in 2026. Old school: • Simple team structure: AEs and more SDRs • Overpaying for outdated tools • Great sales fundamentals New school: • Often have RevOps and GTM engineers • Consistently testing out new tools • More time spent on selling This was made possible ONLY with proper GTM engineering and RevOps infrastructure. Here's the best sales team structure I'm seeing:

@benln The real unlock is relevance. If the ask doesn’t clearly benefit the receiver, even a perfect email gets ignored.

Knowing how to send great cold emails is a superpower. The best ones: • 2-3 sentences, read like a text • have a clear ask • no signature • thoughtfully personalized • punchy subject line • avoid vague requests like "pick your brain" • don't look like AI was used

@jprichardson 24/7 settlement matters, but most capital still values stability over speed in crises. That’s where gold still holds ground.

Bitcoin flipping gold during geopolitical crisis comes as a surprise to no one who understands that 24/7 global settlement with no counterparties always beats waiting for markets to open and vaults to process withdrawals.

The agentic payments space is about to make the same mistake every new fintech category makes. Ship the technology. Assume the trust. Speed to market is being treated as the competitive advantage. It isn't. Trust architecture is. The teams that build it first won't just win users. They'll make every competitor who shipped without it look reckless by comparison.

Using crypto as collateral is bullish, but it also means it’s being absorbed into the same system it was meant to bypass.

@XenBH Credit is the real unlock, but it’s also where things break. Without strong underwriting, onchain credit just becomes faster ways to blow up.

Most people think banks work like a warehouse. You deposit money. They store it. They lend some of it out. It actually works backwards. Banks create money when they issue a loan. The lending itself brings new money into existence. Money and credit are two sides of the same coin. Our entire economy runs on credit. Crypto has done a lot around payments, markets and self custody. All real progress. But credit creation, the mechanism that actually fuels the economy, is still barely explored onchain. Feels like it deserves a lot more attention.

@SHL_eth Curation is leverage now. The feed isn’t about volume anymore, it’s about signal density.

There's never been a better time to curate a list The timeline is full of bad takes, rage bait and people bored out of their minds posting nonsense Curate a list of people who actually add value to your feed People who make you think, teach you something or inspire you to do better Stop doomscrolling through garbage when you could be learning from the best people in your space Your timeline is only as good as the people on it Fix your feed and watch how much better this app becomes

@Tocelot Most “moats” here still depend on the model layer behaving. If capabilities jump enough, even embedded workflows get redefined.

Every era of apps has a defining concept for how you win: mobile was distribution, SaaS was switching costs For AI apps, it’s Minimum Viable Moat: the smallest edge that survives a 3x capability jump from your model provider Foundation model labs aren’t your traditional incumbents. They ship fast and hard. If your product lead evaporates when Opus 5 ships, you don’t have a defensible business What passes the test: - network effects that compound with additional users - workflows too embedded to rip out (vertical AI thrives here) - licensed or proprietary data (80% of the world’s data is private!) - data loops your users create as they use the product (ex. personalization) - brand trust built over many positive interactions with your product if you’re not building toward one or more of these, you’re building on a temporary model capability gap. expect it closed in 3-6 months.

@davidtsocy Speculation is great for acquisition, terrible for retention. Most apps never make it past that handoff.

The next breakout onchain consumer app won’t fight speculation / financialization It’ll use it as the hook to get users in the door, then make them stay with a core mechanic that delivers more joy than just pure trading

@Chris_Orlob The demo works when it reframes the problem, not when it tries to prove the solution. Otherwise you anchor too early

"Don't show the product on the first call" is a LIE. Sometimes a teaser demo OPENS discovery. Show just enough to get their mind spinning. Not too much. Not too little. Use it to spark questions and conversation. Discovery and demos can be happily married together.

@jasonlk v1 is easy now. The bottleneck just moved to ownership and maintenance, which AI doesn’t fully solve yet.

Everyone celebrates when the vibe-coded app hits production. Nobody talks about what happens next. New features. DevOps. Upgrades. Performance at scale. Security patches. That's where most prosumer vibe coding stalls — at v1.3, with a founder who can build but can't maintain. Autonomous 24x7 agents for production, not just vibe'ing your idea, are the solution. And they're coming fast.