JustJerry

@askmape The handoff after the AI summary, when nobody owns the next move.

A growing GTM problem: companies have more customer data than ever, but less shared context AI can sum calls, emails, and notes, but teams still struggle to turn that into coordinated execution What’s currently breaking context flow in your team?

@ibuildthecloud Fancy IDE losing to a hacked-up workflow feels very 2026.

woo hoo, I got agent review working. I'm really like my coding agent. My hacked up thing is more useful to me than any fancy IDE right now. That's nuts. Things are really changing.

@GaurangGujrati GTM has a way of making the repo feel peaceful.

I've been working on my startup for a while now and tech honestly is the easiest part. Figuring out GTM is so hard.

@Rohan27s Customers replying to your emails is a weirdly honest moat.

Building software stopped being a moat a while ago. Anyone can ship now. What actually protects you in 2026: an audience you built before you needed it, customers who reply to your emails, a niche so specific that no funded team would bother the solo founder, the advantage is proximity. Use it.

@AishwaryaDevv Why are you worried about the sky falling down?

AI gonna crash the same way NFTs did.

@adriwtm @nexudotio Mockups get a lot less precious once the agent can open the page and poke it.

@nexudotio nailed the real Codex shift. x.com/nexudotio/stat… Design should not stop at mockups. If Codex can inspect files, test interactions, and critique output, the workflow gets real. This is how design-to-dev stops being handoff theater.

The best tech stack is English

@hiyouga_dev English, I think 🤔

@JustJerry121 What's the best tech stack for cc?

When building a project, choosing the right technology stack is very important. I suggest that you first clarify the product requirements, and then repeatedly communicate with AI about the choice of the technology stack. If the technology stack is wrong, the subsequent workload will be unimaginably amplified.

@aadilbuilds No conversion drop after doubling price is a pretty loud answer.

solo founder tip nobody gives you: price higher than you're comfortable with. I launched at $4.99/mo. felt right. moved to $9.99. same conversion rate. literally doubled revenue by changing one number. your imposter syndrome is not a pricing strategy.

@TMHTechnologies Congrats🎉 keep it going!

Thanks to everyone for connecting! Broke 200 today, huge milestone!

@michaelheredia Support stealing sales time is a very real automation brief.

Most automation pitches are too broad. For a solo founder, start where the day actually breaks: support questions interrupt sales work all day. Then make the agent produce a protected focus time.

@PavithraCharan A product demo or one clear teardown can feel more natural than pretending to be a creator.

Today, most founders are being told they have to build a personal brand online and commit to “founder-led branding”. I know that leaves many founders feeling stuck, overwhelmed, or resistant to even getting started. Not because they don’t think founder-led branding works. But because they are being conditioned to approach it the wrong way. Founder branding today is often presented as: • face reels • creator behaviour • constant posting • high-energy online personas And many thoughtful founders instinctively feel: “This doesn’t sound like me.” They are right. Here’s a reset: The goal is not to create hundreds of identical founder-creators online. The real opportunity is to build differentiated thought leadership that reflects: • the founder’s strengths • communication style • leadership personality • business context • and the brand’s values Some founders are naturally strong at: speaking, writing, frameworks, or strategic commentary. Others are good to showing up in: communities, giving product demos, educating people. Still others, excel at podcasts, interviews, speaking gigs. The point is: Founder-led branding works, it is important. But the founder can choose how they wish to show up in a way that is authentic and aligned to their natural strengths and their business/brand values. Because the goal is not content volume or creator energy. It is trust transfer. That’s best built through showcasing your expertise, skills, perspective and conviction. That’s all your buyers want to see. #Marketasana #FounderBranding #ThoughtLeadership #PersonalBrand #FounderMarketing #MarketingStrategy #LeadershipBranding

@PaulGugAI Tokens per accepted diff is the number I’d want before calling it efficient.

Interesting little example from Jason / Claude Code: HTML generation can burn a surprising chunk of tokens on structural markup. That’s why these workflows need to be judged on actual cost, reliability, and output quality rather than how clever it looks. If the job is mostly repetitive page scaffolding, a cheaper or more targeted approach may be the better trade.

@leonbasinwriter @HarryStebbings Follow me back and I’ll tell you.😜

@JustJerry121 @HarryStebbings Does it connect to any api’s or mcp’s?

I just interviewed a CEO who said three things that blew my mind: 1. We replaced our $600K Salesforce contract with a vibe-coded CRM, built within 3 weeks. 2. We will get rid of 80% of the SaaS we use internally. 3. If Anthropic doubled pricing, we would not change usage in any way.

X is the new moltbook. @moltbook

28 is not too old. Shipping with a birth-year field might be.

@sarthakxyg Per-slice placement is a much saner unit of trust than the whole ticket.

A few months back we were figuring out where to first put an AI coding agent. Do we hand the whole ticket to the agent or do we wait for the next model before letting agents anywhere near production code. We did neither. We pulled the ticket apart instead. A normal engineering ticket is not one job. It is spec gathering, then contracts, then UI, then unit tests, then end-to-end tests, then a database migration, then deploy. Each of those slices has its own reversibility, its own observability, its own blast radius. UI work in a branch you can undo in two minutes; a bad migration on production you cannot. Treating the ticket as one decision forces the same answer onto rows that should not have the same answer. So we made the placement call per slice. UI implementation went first because the worst case was a broken layout in a branch. The first few attempts were mediocre, and that is when the second thing clicked for us. The lever was not the model. It was the substrate the agent was running on: what the IDE could see, which design tool it was wired into, what typed integrations it had to our component library, what rules we had written for the mistakes it kept making. We did not change the model for weeks. We improved the substrate over time, added more layers to it, and a large fraction of the team's UI work moved from human-led to agent-led on the same model that had been mediocre at the start. Full essay sarthakxyg.com/systems/agents…

@theborfeng The “just send pricing” one usually means they’re shopping before the problem is named.

4 things a founder says that tell me they'll fail at email marketing: ❌ "Our list is from 2-3 years ago" ❌ "We don't run ads right now" ❌ "We're hoping email turns things around" ❌ "Just send me your pricing" (before I've even audited their account) Email doesn't fix a business with no momentum.

@THEDEVRELMARKUS I’d start with rough wireframes, then ask AI to critique the hierarchy before styling.

Anyone have a workflow they use for designing websites. I feel like all of mine still give an AI generated look. Post below if you have a recommended tool I should try.