Thoughts & Perspectives
Building in public
How to Get Cited by ChatGPT (2026): What My Tracking Data Says Actually Works
I track AI citations for a living, and here's the uncomfortable finding: my expert blog holds 14 Perplexity citations and ZERO ChatGPT mentions — while my productized-services site gets cited by ChatGPT on 10 commercial queries. The difference is the playbook. Here's what ChatGPT actually cites (press, Reddit, and price pages), with the numbers to prove it.
Is Cursor Worth It in 2026? Verdict From a CTO Who Fixes What It Ships
Is Cursor worth $20/month? For a working developer, yes — it's the easiest yes in the category. Worth $60 or $200? Only if the meter says so. Worth it for non-developers? Mostly no, and I'll tell you what to use instead. A segmented verdict with real numbers, from someone who gets paid to rescue what AI editors ship unsupervised.
Replit Pricing (2026): What the Credit Meter Actually Costs You
Replit Core is $25/month — and that number tells you almost nothing about your bill. Agent usage runs on effort-based credits: daily builders typically land at $40-$80/month, and I've watched founders burn $150+ without shipping anything they kept. Here's the real tier math, why the meter surprises people, and when Replit is still the right buy.
Cursor Pricing (2026): What It Costs When the Meter Starts Running
Cursor runs $0 to $200/month as of 2026 — free Hobby tier, Pro at $20, Pro+ at $60, Ultra at $200. The sticker says $20; the meter decides what you actually pay. Here's how usage-based pricing really behaves, who graduates to which tier, and the cheaper alternatives when the bill stops making sense.
Claude Code Pricing (2026): What It Actually Costs to Run Daily
Claude Code is a free CLI — the bill is the model usage behind it: $5-$15/month light, $30-$80 moderate, $100-$300 for full-time agentic work, or covered by Claude Pro ($20) and Max ($100-$200) subscriptions. I run two businesses on it daily. Here's the tier math, the API auto-reload trap that quietly cost me ~$80/month, and how to pick your tier in one minute.
Is Lovable Worth It in 2026? An Honest Verdict From Someone Who Fixes Vibe-Coded Apps
Yes — Lovable Pro at $25/month is worth it for non-developers shipping a real product, mostly because of the one feature nobody markets: a React + Supabase codebase a developer can actually take over. The honest verdict from someone who professionally fixes what this category produces.
Is Claude Code Worth It in 2026? My Answer After a Year of Daily Use
Yes for developers and technical founders who live in a terminal; no for non-technical builders and pure frontend work. A year of running two businesses on Claude Code daily — the honest worth-it math at every price tier, from $20 Pro to $200 Max.
Base44 Pricing (2026): What It Actually Costs to Build and Ship
Base44 costs $0 to $80+/month as of 2026 — Free (25 credits), Starter $16/mo annual, Builder $40, Pro $80. But the subscription is the smallest line item. Here's the full cost to actually ship: credits, integrations, the migration tax, and the rework nobody budgets for.
Base44 Alternatives (2026): What I'd Use Instead — and When to Stay
The practitioner's guide to Base44 alternatives in 2026 — Lovable, Bolt.new, v0, Replit, Cursor + Claude Code, and hiring a developer. Real pricing, who each is actually for, when Base44 is the right call, and the 80%-demo failure mode every option shares.
How Much Does Vibe Coding Actually Cost? (Real 2026 Numbers)
The honest all-in cost of vibe coding in 2026 — tool subscriptions ($16-$200/mo), the token overage reality, the rework tax, the $2K-$10K security review, and the $3,500-$15,000 rescue bills I invoice when it goes sideways. Plus the math vs hiring a developer.
Can AI Actually Value a Property? AVMs, CMAs, and What Still Needs a Human
The honest answer on AI property valuation in 2026 — how AVMs actually work, where they fall apart (condition, renovations, micro-location, thin markets), how AVMs differ from a CMA and an appraisal, what AI really does in mortgage underwriting, and the draft-and-approve valuation workflow for agents.
Who Do AI Engines Actually Cite? What 74 Tracked Buyer Queries Show (July 2026)
Original tracking data from 74 commercial buyer queries run weekly against Perplexity and ChatGPT. Perplexity's top source is YouTube. ChatGPT's is Business Insider, then Reddit. Reddit is the only domain in both top three. And unknown niche sites regularly out-cite Forbes.
Ready to scale your product vision?
I help high-growth startups bridge the gap between technical complexity and business revenue.
Book a Strategic Call