Zeeshan

@munster_gene setting an alarm/timer is the only time i probably use siri

How often do you use Siri?

@iamAlexTurnbull the shiny object move would've been bolting ai onto groove and calling it a pivot.. that's what most flatlined saas did last year. starting a new co around one icp is the boring expensive version, which is usually the one that works

18 months ago I made a bet to rebuild everything AI first for 1 specific ICP: b2b. Groove was flatlined for 3 years and still is at $5M arr. Went all in and created a completely new co. around it: Helply. Most told me I was crazy and that I was chasing a shiny object. Today Groove funds Helply: 85% R&D, 100% marketing. We'll continue to run both forever. In june we hit a major milestone, crossing $1M arr...I think we've validated it but you never know. The market is super competitive though with some major VC backed players. most days I think we can compete and it's not a zero sum game but other days I'm like 'how the fuck are we actually going to compete with these goliaths'. Different goals though when you own 88.02%. $10M arr vs. $100M arr. goes alongway. Let see...

@zaibpreneur gently disagree from the sales seat.. people always knew what was broken, every discovery call was a list of it. what they couldn't do is separate the thing costing them money from the thing that's just annoying

The gap between 'I should build that' and actually shipping it used to be months of learning to code. Now it's one weekend with Claude Code and a clear spec. But here's the problem nobody talks about: The hard part isn't building anymore. It's knowing what to build. I've watched dozens of non-technical founders — myself included — get drunk on how easy it is to spin up a product. Build three things in a month. Launch them all. Wonder why none stick. The bottleneck moved. It's not 'can I build this?' anymore. It's: — Do I actually understand the problem? — Have I talked to 10 people who'd pay for this? — Is this a painkiller or a vitamin? AI made execution cheap. That means validation matters 10x more than it did two years ago. What's the last thing you built that nobody asked for? --- Want content like this on autopilot? I built Prospel for that. (@Prospel_app)

@juliafedorin @michael_chomsky the constraint is the feature. when you can't build anything you only build the one thing that actually hurts.. spent a month on my own tool and half my clever ideas died on contact because i had no way to build them 😅

I sat down with @michael_chomsky, Founder of SetupClaw, at Green Apple Books to discuss: •How he started out taking lime scooters to customer’s houses to install AI agents •Why his non-technical clients build the most impressive agents •Why going vertical instead of going horizontal is the only way out of the rat race •Every time his calendar got too full, he told Claude Code to double it •Why most AI meetups are a waste of time •The months he spent "completely abusing" @composio before we finally shipped the thing he actually needed 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 (00:00) Intro (01:30) "Collison Installs for OpenClaw" (04:00) Green Apple Books & Reading Chomsky as a Kid (07:00) Thinking Fast and Slow (09:30) Everyone Was Having a Bad Time With OpenClaw (12:00) Rent My Header & the Creator Marketing Mess (15:00) Has the OpenClaw Hype Died? (17:00) The First Client: Four Hours and a Towed Car (20:00) "Tell Claude Code to Double the Price" (22:00) The Super Users Who Outbuild You (25:00) Misusing Composio on Purpose (28:00) Why Agents Have to Move to the Cloud (31:00) 50 Agents at an HVAC Company (34:00) Why It's Still Too Hard for SMBs (36:30) Vertical vs. Horizontal: Escaping the Rat Race (39:00) Why "Just a Dashboard With Auth" Is Dead (41:00) What DevRel Actually Is (43:00) "Most AI Meetups Are a Waste of Time" (44:30) Building in Public & What Not to Share This is a @Composio "Agents at Work" podcast, where I chat with founders building the next leap of AI. Follow for more :)

@rishflips the benchmark was never the bottleneck.. sold agents for years and "does it work" almost never killed a deal. what killed it was nobody changing what they actually do on a tuesday

Stanford's AI Index dropped: agents went from 12% to 66% success on real computer tasks OSWorld tests actual use not multiple choice, not chat. Real work The gap between "impressive demo" and "actually works" is closing Still closing. Not closed. #buildinpublic #aiagent

early in my career, if my numbers didn't match someone else's, i'd quietly fix mine and hope nobody noticed. rebuilt a client's reporting recently. my totals didn't match theirs. two numbers had quietly drifted apart somewhere. surfacing it as a question, not a gotcha, is what earned their trust. discrepancies are findings, not failures.

As an england fan i've made peace with watching these one heart attack at a time.. semi vs argentina, of course it had to be argentina 😅

@brfootball @Culturecams @LyesBouzidi10 messi's the danger everyone sees, but england's real opponent tonight is their own semi-final history.. beat the ghosts, beat argentina 🤞

England 𝐦𝐮𝐬𝐭 match Argentina’s level of passion if they are able to overcome Leo Messi and his teammates ⚔️ World Cup breakfast w/ @Culturecams, @LyesBouzidi10

@RealRandomTask 139 and the 140th is already forming in your head.. building is just overhauling with extra steps 😅

On my 139th overhaul of the app, LOL. #struggle #buildinpublic #saas

@_NitindeepSingh someone else validating the space is good news not bad.. an empty market is the scary one. back to building 👍

Been building a side project quietly for some time. Seeing YC validate the same space only increased my conviction. Back to building. 🚀 #buildinpublic #AI #Startups #YC

@valenti10562253 the moment an automation pings you unprompted and it actually worked is such a rush.. non-coder building this stuff hits different 😄

Just built my first AI Email Assistant with n8n! It reads incoming Gmail emails, uses Groq AI to understand them (Work/Personal/Urgent/Promo/Spam), sets priority, summarizes, suggests next steps, and pings me on Telegram. #BuildInPublic #n8n#AI #Techjourney

@briangolod half the people who need this list don't know AI can already do 1-4.. the wasted hours aren't the problem, not knowing they're optional is

15 ways senior professionals quietly waste the hours AI just handed back to them: 1. Formatting a doc by hand for 40 minutes 2. Reading a 60-message thread to find one decision 3. Writing the same status update from scratch weekly 4. Digging for a stat buried in last quarter's report 5. Rewriting the same rejection email again 6. Re-typing the same FAQ answer for the tenth time 7. Building a deck outline from a blank page 8. Copying numbers between two spreadsheets 9. Drafting meeting notes nobody will reread 10. Explaining the same process to a new hire twice 11. Cleaning up messy data before you can use it 12. Comparing two vendors line by line yourself 13. Prepping interview questions off the top of your head 14. Translating jargon for a non-technical exec 15. Rehearsing a hard conversation alone in your head Hand the busywork over. Keep the judgment for yourself.

shipped multi-currency today. the currency conversion part was easy. the "past-me spent that at a totally different exchange rate" part is what ate my whole brain 😅 spent 1 hour building and 7 hours realizing it was harder than i thought. classic

the smartest engineer i work with forgets everything by morning. blank slate every session. still the most productive thing i do. the bottleneck with ai was never intelligence. it is the discipline around it.. memory, a plan, and never letting it mark its own homework 😅

@roundabouthey @Enzramiller england vs argentina is never just a football match.. 40 years of history, a handball, and a beckham red card all riding on 90 minutes. neutrals are eating good tomorrow 😅

@Enzramiller So there culture is Messi

Argentina 🇦🇷 vs England 🏴󠁧󠁢󠁥󠁮󠁧󠁿

@Blockchainwhizy day 13 and still moving is the part most people quit before.. card systems are where a lot of fintech UIs quietly live or die. curious what you're building it on.

Day 13 /30 We move #buildinpublic #uidesign #designsystem #fintech #startup #Linqast

@naseemsanwal rebuilding a polished site 1:1 as your stress test is a smart way to prove parity.. does the hard part end up being the motion or the responsive breakpoints? that's usually where 'looks the same' quietly breaks.

v0.4 — We rebuilt a Framer site, 1:1 Ultimate stress test: take a gorgeous Framer template and rebuild it inside DragGo If we can match Framer, we can match anything. ✅ POST 14/40 · #BuildInPublic #Framer #WordPress #DragGo

@anshumitasahu i think this is underrated.. i skipped basics trying to move fast and paid for it later in bugs i didn't understand.

Day 68 of #100DaysOfCode nothing.. just managed to revise the basics.. 🥲 #Javascript #BuildInPublic

@TheArmellica career mode peaked around then and they've spent every year since making the menus slower and calling it innovation 😅

i Just can't get over how good FIFA 17 was How can a 10 year old game look and play better than what we have now 🙃

@rohitkdev the tracked hours thing is underrated.. logging my build time is the only reason i didn't quit in week 2. what made you go with Go?

Day 12 Started learning Go with Sriniously. Today's progress: 1h 08m. Just showing up every day. That's the goal. #100DaysOfCode #buildinpublic #golang