In partnership with
Hey,
You've already changed how you use the internet. You probably didn't announce it. You just stopped clicking.
You ask ChatGPT instead of Googling. When you do Google, the AI answer at the top is usually enough. The numbers say you're the majority now:
🔎 ~60% of all Google searches now end without a single click. For news questions, it's 69%.
📰 Traffic from Google to the world's websites fell 33% in a year. Some sites lost up to 89%.
🤖 Google's new Search is, in its own words, "AI search through and through" — answers, not links.
Convenient, right? It is. But three things happened this week that change the deal you're getting:
One — the ads arrived. ChatGPT ads officially launched in the UK. The free answer you read is about to be a monetized surface, same as search results were. The line between "the answer" and "the sponsored answer" is about to get blurry.
Two — the sources started walking away. Reuters and Time switched to blocking all AI bots this week — whitelist-only. The Economist is building a two-track internet: one version for humans, one for machines. Translation: the AIs you rely on may be answering from older, thinner source material than you think.
Three — nobody's checking. When 6 in 10 questions end at the AI's answer, the answer never gets stress-tested by a click.
So this issue is your survival kit as a reader: how to make any AI show its receipts, plus the everyday tools that earn their spot in an AI-first life. Skim what you need.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
AI Agents Are Reading Your Docs. Are You Ready?
Last month, 48% of visitors to documentation sites across Mintlify were AI agents, not humans.
Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.
This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.
Your docs aren't just helping users anymore. They're your product's first interview with the machines deciding whether to recommend you.
That means: clear schema markup so agents can parse your content, real benchmarks instead of marketing fluff, open endpoints agents can actually test, and honest comparisons that emphasize strengths without hype.
Mintlify powers documentation for over 20,000 companies, reaching 100M+ people every year. We just raised a $45M Series B led by @a16z and @SalesforceVC to build the knowledge layer for the agent era.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔥 TOOL OF THE WEEK — Consensus
If the theme of the week is "AI answers without receipts," Consensus is the counter-example: an AI search engine whose entire product is the receipts. Ask it a question — does creatine actually work, is screen time harming kids, does intermittent fasting do anything — and it answers from 200M+ peer-reviewed papers, with every claim cited and a "Consensus Meter" showing how much the research actually agrees.
Three things to know:
It's built for the questions you currently ask ChatGPT and shouldn't. Health claims, nutrition, sleep, parenting, supplements, psychology — the categories where a confident-sounding wrong answer costs you the most. Consensus won't speculate; if the literature is split, the meter shows you the split. That honesty is the product.
The free tier is real. 10 Pro Analyses, 10 Study Snapshots, and 10 Ask Paper messages every month — refreshing monthly, not a one-time trial. Premium is $8.99/mo (~$108/yr) for unlimited analyses; Pro at $15/mo adds Deep Searches for serious literature digging. Students get 40% off — Premium lands around $5.39/mo, the best student deal in the category.
It pairs with your chatbot, not against it. The workflow that works: ChatGPT/Claude for thinking, drafting, planning — Consensus for any factual claim you're about to act on or repeat to another human. Two tools, two jobs.
What it's NOT good for: news, current events, opinions, or anything outside published research. It can't tell you who won last night's game or whether a movie is good. It tells you what science says — that's the whole job, and it does it better than anything else.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚡ THREE QUICK HITS — the everyday stack
Reclaim — the calendar AI for people whose week gets eaten alive. Connect it to Google Calendar, tell it what matters — gym 3×, deep work every morning, lunch that doesn't get scheduled over — and it defends those blocks automatically, moving them around incoming meetings instead of letting meetings delete them. The free Lite plan covers one habit and basic smart scheduling — enough to feel the magic. Paid starts around $8–10/user/mo (annual vs monthly billing) for unlimited habits, task sync with Todoist/ClickUp/Asana, and buffer time between calls. Honest caveat: if you only have 2-3 meetings a week, you don't need this — it shines when your calendar is the battlefield.
→ Try Reclaim free
Murf — if you've ever wanted a real voiceover for a YouTube video, a course, a TikTok, or a slideshow without recording yourself fifteen times: this is the tool. 200+ studio-grade AI voices in 30+ languages, a browser editor where you fix pronunciation and emphasis per word, and direct text-to-speech that sounds like a hired narrator. Creator is $19/mo on annual ($29 monthly) with full commercial rights — meaning you can monetize what you make. The free tier is preview-only (10 minutes, no downloads), so treat it as a test drive. Caveat: cloning your own voice is locked to higher tiers — if that's your main goal, this isn't your pick.
→ Try Murf
Miro — the infinite whiteboard that became the default place to think visually. Mind-maps, trip plans, project boards, brainstorms with friends or teammates — all on one canvas, now with AI that turns a messy sticky-note dump into structured diagrams and summaries. The free plan gives you 3 boards forever, no card required, which is genuinely enough for personal use. It earns a spot here for a simple reason: when AI gives you instant answers, the scarce skill becomes organizing your own thinking — and a canvas beats a chat window for that.
→ Try Miro free
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✍️ PROMPT OF THE WEEK — The Source Check
The single highest-value habit you can build this year. After any AI answer you're about to act on — health, money, legal, a fact you'll repeat at dinner — paste this as the follow-up:
Before I act on that answer, audit yourself:
1. Sources. What specific sources is this based on? Name them. If you can't name real sources, say so plainly.
2. Confidence. Rate your confidence 1–10 for each claim you made, and tell me which single claim is most likely to be wrong.
3. Freshness. Could this have changed recently? What's the date of the newest information you're working from on this topic?
4. The other side. What would a well-informed person who disagrees with this answer say? Steelman them.
5. Verification. Give me the one search I should run or the one source I should read to verify this independently.
Why it works: AIs are trained to sound certain, not to be certain — and now that publishers are blocking the bots, the gap between confidence and currency is widening. This prompt doesn't make the AI smarter; it makes the AI honest about its own limits, which is the next best thing. Takes ten seconds. Catches the expensive mistakes. Steal it, save it, use it forever.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚖️ COMPARE CORNER — Gamma vs Beautiful.ai
The most common "which one?" question in our inbox: the two AI presentation tools, for the deck you have to give next week — class, work, side project, wedding speech with slides (it happens).
Gamma — fastest idea-to-deck on the market. Type a prompt, get a full presentation, edit like a doc. Free plan with 400 one-time credits (roughly 10 decks) — no card needed. Plus is $8/mo annual ($10 monthly): removes branding, unlocks PowerPoint export. Pro at $15/mo annual adds unlimited generation. Best for: students, one-off decks, anyone who wants free to actually mean free, and people who'd rather edit text than wrestle layouts.
Beautiful.ai — the design-enforcement engine. Its Smart Slides system makes it physically impossible to produce an ugly slide: every layout auto-balances as you add content. Pro is $12/mo billed annually; there's also a $45 one-off "Single Presentation" option — genuinely clever if you present twice a year. No free plan, and heads up: the 14-day trial requires a card and auto-charges if you don't cancel — set a reminder. Students with a .edu email get a free year. Best for: people who present to clients or bosses where polish is the point, and recurring presenters who want a locked-in visual identity.
The honest take: Gamma for speed and zero-cost starts; Beautiful.ai for polish and design discipline. If you present occasionally — Gamma free tier, done. If your slides are how you get judged professionally — Beautiful.ai's $12/mo pays for itself the first time someone asks "who designed this?" And if you literally have one big presentation this year, the $45 single-deck option is the sleeper pick nobody talks about.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📖 FROM THE TOOLCHASE BLOG
Three reads that pair with this issue:
The tools that are genuinely free — not "free trial," not "free until you need the one feature that matters." Our most-shared list, updated for 2026.
Perplexity, ChatGPT Search, and the rest — where each one pulls its answers from, which ones cite sources properly, and which to trust for what.
The phone-first picks — what's actually worth a spot on your home screen this year.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Gauntlet AI: A fellowship for experienced software engineers
AI changed what senior engineering means. Gauntlet is built for engineers who want to lead that change — not watch it. The most effective way to become AI-first. No tuition. No tutorials. Just the work. Apply now.
Must be a US citizen to qualify.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📌 ONE EDITORIAL TAKE
Here's the trade you're being offered, spelled out: infinite free answers, in exchange for not knowing where they came from.
For most questions, that trade is fine. What's the capital of Slovenia, how long to boil an egg, summarize this email — provenance doesn't matter, convenience wins, take the deal.
But a slice of your questions are different. The medical one. The financial one. The "should I sign this" one. The fact you're about to repeat to someone who trusts you. For those, the AI's confidence is not evidence — and as of this week, with publishers pulling their content and ads entering the answer box, the gap between sounding right and being right is officially widening.
The fix isn't abandoning AI. It's a two-speed habit: fast mode for the 90% where being roughly right is fine, receipt mode for the 10% where it isn't. The Source Check prompt is receipt mode. Consensus is receipt mode. Paying for — or at least keeping — one human source you actually trust is receipt mode.
The web is splitting into a track for humans and a track for machines. You'll live on both. The skill of the next decade is knowing which track your current question belongs on.
That's it for this Wednesday.
Tell me: what's the last thing an AI told you that turned out to be flat wrong? Hit reply — we're collecting these, and the best ones make it into a future issue (anonymously, promise).
Talk next week,
Emre
Founder, ToolChase
P.S. Building an AI tool? Submit it for a free editorial review →
P.P.S. Forwarded this email? Subscribe here →