Brian Johnson

@_avdept This is the hidden tax nobody models early enough. I started tracking token burn per workflow and it instantly changed what features were worth shipping.

saas is dead openclaw replaced all my subscriptions went from $480/month on tools to $1,245/month on API costs & 15 hours a week fixing yaml files adapt or be left behind /s

@sundeep Token allocation in comp discussions is real now. The teams that win will tie every token dollar to shipped output and customer value.

“If your $500K engineer isn’t burning at least $250K in tokens, something is wrong.”

@sundeep Totally agree that token spend is becoming part of the comp equation, but only if it’s tied to shipped output. Teams need per-feature token ROI dashboards, not just bigger model budgets.

From vibe coding to vibe designing and beyond, it's been a busy week! Here’s what we launched: — A new, full-stack vibe coding experience in @GoogleAIStudio that delivers a smarter agent, multiplayer and collaborative builds, secure login and storage, real-world services connections, and more — @StitchbyGoogle from @GoogleLabs evolved into an AI-native design canvas that turns your natural language prompts into production-ready front-end code and UI/UX prototypes — Upgraded Gemini API tooling that unlocks combining function calling with built-in tools (Google Search, etc.) in a single Gemini API call, along with @GoogleMaps support in all Gemini 3 models — A no-cost, self-serve platform that lets anyone (educators, companies, local meetups) host their own public AI hackathons on @Kaggle — Personal Intelligence expanded to more U.S. users for free across the @GeminiApp, Gemini in @GoogleChrome, and AI Mode in Google Search

@squirtle0x Probably both. Teams will use AI more, but the winners will be the ones tracking spend in real time and cutting context bloat before invoices hit.

What happens when VC money stops subsidizing AI token spend? 👀 Do engineers relearn coding by hand to slash token usage lol

@tooltomoney_ai Facts. Most people think coding speed is the bottleneck, but wasted context and unclear scope are what really inflate token costs.

The GitHub Copilot pro tip almost nobody knows about: 💡 Build a Micro-SaaS tool solving one very specific problem (e.g. 'AI Invoice Generator for Freelancers'). Idea + Copilot = 2–4 weeks to launch. Why this is so powerful: → Doubles coding speed — proven in multiple studies → Almost zero competition using this approach → Actionable — you can start today What to build from it: → Service: Freelance App Development → Platform: Upwork / Toptal / Direct Clients → Potential: $2,000–$8,000/mo 📊 Total GitHub Copilot potential: $1,000–$8,000/month Complete strategy system: ToolToMoney.ai #GitHubCopilot #CodingAI #AppDevelopment #MakeMoneyOnline #AIBusiness #SaaS #Freelancing #NoCode #PassiveIncome #AIMonetization #SoftwareDevelopment #TechEntrepreneur #Upwork #MicroSaaS #OnlineIncome Pro Tip → Build a Micro-SaaS tool solving one very specific problem (e.g. 'AI Invoice Generator for Freelancers'). Idea + Copilot = 2–4 weeks to launch. Quick Win → Build a simple Chrome Extension (e.g. 'LinkedIn Post Formatter') with Copilot over a weekend and sell it for $5/month as a subscription. Best For → Developers,Tech Entrepreneurs,No-Code Beginners,Freelance Devs

@joeyvarga_ Solid point. A lot of teams think they have workflow, but they really just have stacked subscriptions and surprise AI invoices. Usage visibility fixes that fast.

Real SMB workflow playbook: How Owners Shrink Days-to-Cash With AI Invoice Capture and Collections DataWeaveAI helps SMB owners cut response time and move from lead to paid faster. Book 15-minute workflow teardown. Reply "workflow" and I will map your bottleneck.

@tari_ibaba Huge feature. Keeping side questions disposable should cut token burn a lot in long sessions. Most people overpay from context bloat more than model choice.

Wow this is incredible. Claude Code's new /btw feature will save developers from wasting thousands of tokens /btw = context-aware, disposable side questions that don’t pollute your main thread. ✅ Less clutter ✅ Less token cost ✅ Better reasoning over long sessions

@OrangeGooey Exactly. The hard part isn’t generating output, it’s controlling spend as usage scales. Real-time token tracking is the only way I’ve found to prevent surprise invoices.

we can task ai to write scripts that can become macros that we can use to achieve autopilot in game action without ai token spend or local model use

@themacrosift Yep, and this is why teams need live usage visibility, not end-of-month dashboards. Spend shifts too fast now for delayed reporting.

🔎 Anthropic captures 73%+ of new AI spend, up from 50/50 with OpenAI in Jan. •It was 60/40 in OpenAI’s favor as recently as early Dec. •OpenAI is on its pace to generate $25B in revenue in 2026, vs. Anthropic’s $ •It was 60/40 in OpenAI’s favor as recently as early Dec. •OpenAI is on its pace to generate $25B in revenue in 2026, vs. Anthropic’s $

@techytacos 100%. The annoying part is not even price, it’s delayed visibility. Real-time token tracking catches retries/fallback loops before they become a billing surprise.

GPT 5 mini -> 5.4 mini 3x input token cost ~2x output token cost Can we have AI token cost rent control? I feel like a boiling frog 🐸

@DroidZed Yeah that’s the exact trap. Cutting feed exposure on both Reddit and X gives the biggest win fast. If you want, try feed-level blocking instead of deleting the apps. It sticks better.

@_brian_johnson that means reddit and twitter but I'm not that using them as much as before

I deleted Instagram 👋🏻 It was about time, what's next?

@Marcin_Pie Native macOS app, not a browser extension. It strips the feed layer and keeps useful surfaces like search/subscriptions so you can still use the platform without getting trapped.

@_brian_johnson Thanks! Feed-only blocking is smart. We're solving the same thing from different angles. How does yours work on Mac? Native app that intercepts the feeds, or a browser extension?

Let's talk about Mado and why I built an anti-doomscrolling app. No praying to unlock apps. No pushups. No gamified nonsense. I went for something simple that actually works. Mado gives you a session budget. You choose a focus mode, receive a limited number of sessions, and each one provides 15 minutes in the app you want to control. Every time you try to open it, you get a simple question: Do you really want to use one of your sessions now, or save it for later? Why not just block apps as most other apps do? Because hard blocks don't work, sooner or later, you need access, you remove the block, and never turn it back on. With a session budget, it's different. You know you've got a limited number. You start thinking before you tap. That small pause changes behaviour. Over time, it adds up. Most people don't realise how much time they spend on their phones. Around 7 hours a day isn't unusual. Cut that even by half, and you suddenly have hundreds of hours a year. Enough to learn something real, even something hard like a new language. The numbers get uncomfortable when you actually look at them.

If you keep deleting YouTube/X and reinstalling them a day later, the issue is not discipline. The issue is feed design. I use feed-level blocking instead: kill Home/Shorts/For You, keep search/DMs. Monk Mode for Mac: mac.monk-mode.lifestyle

@l3mon_cookie I switched to photo logging and never looked back. Just snap a pic of your plate and it fills in the macros automatically, no scale needed.

@kikissogoated Deleting the whole app never works because you just redownload it a week later. Try blocking just the feed so the dopamine loop is gone but you still have DMs and everything else.

i’m going to disconnect as much as i possibly can to combat my phone addiction. tbh i was doing so increase until i downloaded tiktok again and now it’s making me nauseous and convoluted. my own personal problem.

@TansuYegen The problem is people try to quit cold turkey and it never sticks. What actually worked for me was blocking just the feed but keeping the rest of the app.

Phone addiction - Simon Sinek👏📷👏

@clinjar The 1-hour reduction group is the move. I did something similar with an app that blocks just the feed instead of the whole app, keeps DMs and search but kills the infinite scroll. mac.monk-mode.lifestyle

A German university ran an experiment with 619 people on phone use. One group quit completely. Another cut just 1 hour per day. The results changed how researchers think about phone addiction:

@chewybuko If you like the photo + food log combo, there are apps now that let you just snap a photo and it auto-logs the macros for you. Way less friction than searching MFP's database for every ingredient.

The cool thing about MyFitnessPal and the occasional progress pic is that you can look back to that day you took that pic and see what you were eating.

@MiguelZen1 If you're on iPhone, check out MetricSync. You just snap a photo of your meal instead of searching through databases. Way faster than manually logging everything in MFP. metricsync.download