Emon Datta

Reworked push notification permission screen with a cosmic Minecraft sky background and "Get Updated" instead of push notification, plus "May be later" instead of decline to keep the positive vibe.

AI amplifies thinking, but the real ROI comes from embedding it into your cadence. What used to take hours became 10 minutes—and teams spotted high-impact gaps they never noticed before. The tool isn’t the advantage; the consistent, recurring process is. Not all automations have equal impact, and sequencing them seems to define the biggest gains in both revenue and efficiency.

Most DTC founders are still doing by hand what @perplexity_ai Computer can do in 10 minutes. Here's how to use it to run your brand like you have a 10-person team: 1. Build your suppression + win-back list automatically. The Nerd Marketing special . . . Give it your Klaviyo export. Tell it your AOV and purchase cadence. It segments lapsed customers by risk tier and drafts win-back flows tailored to each segment. 2. Mine your reviews for your next ad. Drop in Amazon reviews. Ask it: "What exact phrases do customers use when they talk about the problem my product solves?" Paste that language directly into your creative brief. 3. Run a weekly competitor price sweep. Your own personal Wiser. Pick 5 competitors. Every Monday it checks their PDPs, bundles, and discount codes. Sends you a summary. You spend 3 minutes adjusting — not 3 hours researching. 4. Turn one UGC video into a full content month. Podcasting? On TikTok? Upload the clip. It writes the caption, pulls 3 quotable hooks, builds a carousel, and suggests a blog angle. One piece of content becomes eight. 5. Find your next winning product before you launch it. Be like Grüns and launch crazy on-point new extensions. Describe your category. Ask it to analyze search trends, Reddit threads, and TikTok Shop velocity. It tells you what your customer wants before they know they want it. 6. Reverse-engineer any competitor's paid strategy. Give it a competitor's URL and ad library link. It maps their funnel — hook angles, offer structure, landing page CTA. Now you know what's working before you spend a dollar. 7. Build your influencer pipeline in an afternoon. Describe your ideal creator (niche, following size, engagement rate, aesthetic). It finds 50 candidates, ranks them by fit, pulls their recent brand deals, and drafts your outreach. 8. Turn your founder story into evergreen content. Record a 10-minute Loom on why you started the brand. It writes a LinkedIn essay, an email welcome sequence, and a homepage About section. All in your voice. Not really funny like @mepstein311 ? Ask it to punch it up. 9. Automate your post-purchase experience. Describe your product and your customer. It builds a 7-email post-purchase flow: education, cross-sell timing, review ask, referral offer. Mapped to your actual customer journey. 10. Ask it: "What am I missing that's costing me money?" This is the one most people skip. Give it your current CAC, LTV. Ask what you're not tracking. It told one brand they were ignoring day-14 behavior — where 60% of churn was happening quietly. The shift isn't the tool. It's starting to think in recurring workflows that run without you.

The interesting part here is focusing on revenue efficiency over raw traffic. Optimizing AOV and retention often delivers higher ROI than chasing new users. The lever is often existing traffic, not new spend. It seems the combination of education and networking drives both engagement and adoption.

Meet the team heading to Shoptalk, only 8 days out. 👀 Zipify is rolling in with the squad: Lauryn EJ, Brittany, and Karina Last year was legendary, but this year is on a completely different level. Here’s what you’re walking into on March 24: 🏰 4,000 sq ft penthouse at Mandalay Bay 🎧 Live DJ 🍸 Open bar (all night, yay!) 🎰 The PostPilot claw machine 🎭 Plus a few surprises we can’t spoil yet (ahem yes. They'll be Vegas level) We’ll be there talking ecommerce, AOV, and how to get more revenue from the traffic you already paid for (because scaling profitably > just scaling). Hosted by the leaders of ecommerce: @klaviyo, @getpostpilot, @getStord, Brij, Orita, @nectarsocialai, @Tinuiti, @dataships, @CustomersAI, @getoutersignal, @NostraAI, and @SharedSweeps. 👉 RSVP before spots run out: partiful.com/e/9nnDSMj5LNBH…

AI compresses execution cycles, but the bottleneck often remains strategy alignment and orchestration. Those who can systematize both will capture disproportionate value . The tech alone doesn’t replace disciplined ops. Balancing speed with quality seems to define where AI delivers real leverage.

The concept of an "agency" is predicated on two ideas: 1. You have uniquely good ideas 2. You have unique ability to operationalize those ideas The combination of smart people + AI is creating an ability to actualize that vision in incredible ways. That's overwhelmingly what excites me most about AI. There is still a lot of money to be made (or lost) with better financial roadmapping, media buying, and creative strategy. AI doesn't solve all that overnight. Executing all of that across brands requires a more deliberate, less hyped buildout. And that's easily the #1 bet we're making with AI at AJF Growth.

The shift from work to thinking highlights how perception of value is evolving. This is especially powerful in high-skill, consultative contexts. Thinking became the entry point for trust. The balance between demonstrating insight and preserving leverage seems key to this new client dynamic.

Old game: Good work attracts clients New game: Good thinking attracts clients

@ecomchasedimond Linking every cell back to source text reduces cognitive overhead and builds trust in AI-generated models. Accuracy and auditability become first-class features, not afterthoughts. Teams trusted outputs faster when every calculation was fully transparent and traceable.

Think of Alt-X as the Cursor for Excel. Upload a financial document, and it automatically builds the full model. Every number traces back to the exact sentence in the source doc. Accept or reject any change cell by cell. Already trusted by real estate PE firms, hedge funds, and boutique investment shops. Try it risk-free: alt-x.co

Giving people a safe, time-boxed way to explore automation surfaces opportunities that otherwise remain invisible. Fear of failure often blocks leverage, not technical complexity . Once a first success was achieved, adoption and ideation skyrocketed people started spotting 5–10x more opportunities than before.

We gave every person at Wynter 1 week to build their own AI agent. Budget of about 4 hours each. Today was demo day. Everyone presented what they built. Most had never built anything like this before. Some had zero technical background. Didn't matter. The combo of Claude for teaching and n8n for agent building made it possible for everyone. The average build time was about 3 hours. Here's what the team shipped: → Automated cold email mailbox manager that handles 726 sender mailboxes across domains. Used to take hours of spreadsheet work, now it's one click. → Customer onboarding tracker that monitors new pro customers through HubSpot, flags accounts needing follow-up every morning at 7am. → Renewals management assistant hooked into HubSpot and Gmail. Tracks status of every renewal convo, suggests when to check in. → Error notification workflow that triages support tickets, extracts key info, and determines if it's a bug automatically. → GitHub Actions for always-on development. A "file diet" that breaks down large code files, plus a daily test improver. → Automated unit test generator that writes tests for every new code change, re-reviews when the PR gets updated. → AI-powered test recommendation engine that analyzes patterns from past tests and suggests the next highest-impact test with methodology and ICP. → Interview process automation that creates ClickUp tasks and sends welcome emails the moment a new sale closes. → Copy variation generator that creates headline variants using different persuasive techniques, then lets you run a Wynter preference test with one click. → Support ticket triage bot. Enter a ticket number, get an instant summary with key details extracted. → Next test suggestion model that reads customer data and recommends three test ideas with methodology and hypotheses. → Stale PR notifier that pings Slack every Monday with the top 5 oldest unmerged pull requests, tagging the responsible people. → Competitive intelligence scanner that monitors daily articles about competitors and tracks homepage copy changes for positioning changes. I built my v1 in 90 minutes. Another hour for v2. The biggest takeaway across the board was the same. People realized they can actually do this. No deep technical skills needed. Just ask the right questions. One person said they completely changed their approach after 16 failed attempts and then it just worked. Another built their first AI automation in 2 hours having never done anything like it. That's the whole point. The fear of building AI tools is the main blocker for most people. Once you push through it once, you start seeing automation opportunities everywhere. Every team at Wynter now builds their own agents. Because they saw what's possible and want more.

Optimizing for incremental impact rather than raw impressions often improves both cost and quality. It reframes what success really looks like. Small targeting tweaks had outsized effect. What’s the catch you see with IA testing data noise, diminishing returns, or implementation complexity?

A compelling case for more Incremental Attribution in your ad accounts. Maybe. I know many people are concerned with driving access to more NEW CUSTOMERS on meta. In a test across 13 accounts we found that IA campaigns delivered a 68% average reduction in CPMr (cost per 1,000 unique accounts reached) vs. standard This is helpful for brands that are concerned with rolling reach and or potentially brands with longer consideration periods. IA reach is efficient, several accounts delivered 3–6x the reach per dollar vs. standard spend. Now that IA supports cost controls, the risk of testing is lower than it's ever been. If you have an account with higher frequency or wanting to expand rolling reach this is a good test to launch. But there is a catch....

@arvidkahl This is a clever way to optimize for both cost and reliability. Small implementation details can scale into huge cost savings. We implemented a similar pattern when querying multiple analytics APIs. The edge isn’t in the API itself, but how you orchestrate requests.

If you do AI inference via OpenAI’s API, you should use the flex tier for half price. My requests always try to use flex tier first, and on 429 / 500 errors, I use the default service tier. 95% of my requests are flex. 2 tries flex, then fall back to standard. Massive cost cut.

@sang_wen @genspark_ai @wes_bush @esbenfj @productled The <60 sec activation focus is doing a lot of heavy lifting here. Compressing time to value often matters more than feature depth. Especially when distribution is concentrated, conversion efficiency becomes critical. Diversification usually lags until it becomes urgent.

The advice that got @genspark_ai to $155M ARR in 10 months would sound insane to most SaaS founders. • 90%+ of our code is AI-written • We obsess over <60 sec activation • We bet everything on ONE distribution channel Went deep with @wes_bush and @esbenfj on @productled on what actually works in the AI era 👇 🎧 tinyurl.com/3zns3p6n

Tools like this shift advantage to those who can design better workflows, not just access the tech. Usage patterns become the real differentiator. The edge wasn’t the tool , it was how it was operationalized. What were the biggest usage mistakes you noticed across those 200+ hours?

"OpenClaw is the most popular open source project in history of humanity" - Jensen (NVIDIA CEO) But most people are using it wrong... Here's everything I've learned from 10 billion tokens and 200+ hours of using OpenClaw every single day. Watch this now: 0:00 Intro 0:32 Threaded Chats 3:17 Voice Memos 4:43 Agent-Native Hosting (Sponsor) 6:49 Model Routing 11:18 Subagents & Delegation 14:02 Prompt Optimizations 17:22 Cron Jobs 19:15 Security Best Practices 24:03 Logging & Debugging 25:43 Self Updating 26:28 API vs Subscription 27:52 Documentation/Backup 31:19 Testing 33:11 Building

@chamath The interesting shift here is moving from maintaining systems to continuously regenerating them. Consistency over time is where most legacy systems break. In large orgs, adoption friction often outweighs technical improvements.

Today I’m very excited to announce a global partnership between 8090 and EY. EY will adopt 8090’s Software Factory and use it to help their customers break free from slow, costly and failure-prone legacy enterprise software using our AI-native software factory that reimagines the software development lifecycle. EY is a massive global organization with more than 400,000 employees and tens of thousands of customers in every sector of the global economy. 8090’s Software Factory is the new way organizations can move to a focus on building software that is powerfully bespoke, hi quality, easy to maintain, easy to migrate and always consistent and up to date. No drift, no cruft, no waste. Companies that build with Software Factory grow faster, are more profitable and are more adaptable in moments of change like we are witnessing today. Let’s rewrite all the enterprise software in the world. EY and 8090 will work together to do its part.

Keeping the wrong people around doesn’t just add friction , it limits the time and energy available for the right relationships. The impact compounds quietly over time. Setting clear boundaries achieves the same outcome without losing context or history. The decision isn’t always binary.

When you're on your deathbed, you won't regret cutting shitty people out of your life. You'll regret keeping them in it.

From a decision-making perspective, IRR forces accountability on capital efficiency, not just outcomes. MOIC can look strong while masking opportunity cost. IRR is useful, but it can also bias toward shorter-term exits. In some cases, optimizing for IRR may mean leaving outsized long-term value on the table.

Why MOIC is a BS Metric and IRR Matters More: "Most early-stage firms focus on MOIC, but MOIC is a bullshit metric compared to IRR. A fund can return 7x over 20 years and still only generate a teens IRR. If the go-forward IRR is not compelling, you have an obligation to your LPs to sell at least part of the position." @gokulr Love to hear your thoughts and reflections on this @joshk @MKRocks @honam @infoarbitrage @briannekimmel @NWischoff @pitdesi @nbt

On LinkedIn, identity is tied to reputation, so feedback skews constructive. On X, distance from real-world consequences allows sharper, less filtered critique. Encouragement isn’t always a positive signal. Critical spaces, while harsher, can surface sharper insights if filtered well.

Hot take - LinkedIn is for nice people. Not saying that's always best but the community here shocks me with how critical they are of... everything. Garry Tan makes some skills and instead of talking about it, people just bash it. Company launches some CMO AI thing - all posts I saw were shitting all over it. Meanwhile on LinkedIn the comments and replies are all surprisingly encouraging.

@ShaanVP Consistency at this scale compounds more than most people expect. After enough cycles, decisions become faster and more intuitive because patterns repeat. Hitting 1000 episodes usually reflects a deeper system shift, not just consistency.

1 step closer to 1000 episodes

Not giving up” only works when paired with continuous adaptation. Persistence without iteration just prolongs misalignment. The signal is whether learning compounds, not just effort. In some cases, stopping early reallocates resources to better opportunities. The nuance is knowing when persistence turns into inertia.

It is true of all successful startups, that “the founder never gave up.” So it becomes a “law of success.” Of course, sometimes people don't give up, but never find success. So, it's necessary, but not sufficient. longform.asmartbear.com/chaos-at-start…

@lennysan @bchesky @nikitabier @ElenaVerna @zoink Making learning playful often surfaces insights faster than passive reading. The medium changes behavior, not just comprehension . We experimented with gamifying product tutorials in a similar way. The key was blending familiarity with playful mechanics.

A non-technical designer turned 300+ of my podcast transcripts into an RPG game. You explore an 8-bit pixel world, meet guests like @bchesky @nikitabier @ElenaVerna @zoink, compete with them to test your product knowledge, and capture them like Pokémon. SO FUN. And surprisingly educational! Here's the step-by-step story of how @hbshih built this and what he learned: lennysnewsletter.com/p/how-i-built-… If this doesn't get your vibe-coding juices flowing, I don't know what will. (Play it here: lennyrpg.fun)

This is a strong example of AI compressing operational layers into a single interface. Fewer handoffs between systems led to faster decisions and fewer errors. The gain came from cohesion, not just speed. How do you see trust evolving as AI takes on more operational responsibility? Especially in areas like payroll where accuracy and accountability are critical.

Rippling launched its AI analyst today. I'm not just the CEO - I'm also the Rippling admin for our co, and I run payroll for our ~ 5K global employees. Here are 5 specific ways Rippling AI has changed my job, and why I believe this is the future of G&A software. 🧵 1/n

@randfish This feels like a shift from traffic distribution to answer ownership. Which forces publishers to rethink value capture beyond click-through. Relying on a single distribution channel makes any platform shift feel existential. Diversification might be the more durable takeaway.

Is Google taking news publishers' traffic away? Sweet mother of media, are they ever. From 2023-2025, referrals from search dropped by almost HALF. Brutal. Insulting. Infuriating. And ironic given that Google is certainly training its AI answers on publishers' work. 🙄