ChatGPT Just Started Showing Ads, and Your Customers Are Already Seeing Them

Five stories today, and they all point the same direction: AI is racing to get into every corner of how you work, and it is getting cheaper, and in one case free, on the way in. The headline change lands right where your customers already are. Let's get into it.

OpenAI or Anthropic? Which model leads the next benchmark? Which company ships the next major breakthrough?

If you follow AI closely, you already have opinions on where the industry is headed. Kalshi lets you trade on real-world AI and technology events, with markets that move as models launch, benchmarks drop, and announcements happen.

The people who follow this space most closely often see the story developing before everyone else. Put that knowledge to work and trade what you think happens next.

Trade $25, Get Up to $500

_{Bonus credit varies from $15 to $500. Terms apply.}

On July 22, OpenAI began testing ads inside ChatGPT and opened a self-serve "Advertise in ChatGPT" manager that lets brands target people based on what they are chatting about, not just the keywords they type, according to OpenAI. Sponsored blocks appear below ChatGPT's answers, and early advertisers include Best Buy, Lowe's, and VistaPrint. For now the ads show only to adult users on the free and Go plans, while Plus, Pro, and Business stay ad-free. OpenAI paired the move with a push to get small businesses onto ChatGPT and a $30 billion data center in Georgia to power the whole thing.

Why it matters to you: The single most-used AI tool on earth just became an ad platform, which means the place your customers go to ask "what should I buy" is now a place your competitors can pay to show up. Search ads reshaped small business marketing for twenty years. This is the early version of that same shift, moving from Google's results page into a chat window.

What to do about it: Do not buy anything yet, just go ask ChatGPT a question a customer would ask about your industry and see whose ad shows up next to the answer.

Google released three new Gemini models on the same day, splitting everyday AI work from specialized jobs, according to Google. The main workhorse, Gemini 3.6 Flash, gets better at coding and document analysis while using 17% fewer tokens, the units AI usage is billed in, which makes it cheaper to run. A second model, Flash-Lite, is built for speed and high-volume jobs at about $0.30 per million input tokens. A third, locked-down model is reserved for governments and vetted partners to hunt security holes.

Why it matters to you: The interesting part is not the models, it is the strategy. Google is betting that a cheap, fast model that is good enough beats an expensive genius model for most real work. That is good news for your budget, because the AI features baked into the tools you already pay for get cheaper to deliver, and that pressure eventually shows up in pricing.

Canva rolled out Code 2.0 to every user, including free accounts, letting anyone build a working website, app, or interactive page just by describing it in plain English, according to VentureBeat. Once it generates something, you can edit it visually with the same drag-and-drop you already use for graphics, swapping colors, fonts, and images, then publish it as a live site. Canva says the tool is 75% faster than the previous version and now ships with more than 50 templates for interactive pages.

Why it matters to you: A landing page for a promotion, a simple booking page, a quick interactive quiz to capture leads, these used to mean hiring someone or wrestling with a website builder. If you already use Canva for your graphics, you can now try building the page itself without touching code or paying a developer.

What to do about it: Next time you need a one-off page, describe it to Canva Code before you go pay someone to build it.

Anthropic added a feature that lets its Claude assistant learn a task by watching you perform it, according to PCMag. Instead of typing out every step, you record your screen and talk through what you are doing, and Claude turns the demonstration into a reusable skill it can run again later. The feature is rolling out to Pro, Max, and Team subscribers inside the desktop app.

Why it matters to you: The hardest part of handing a repetitive task to AI has always been explaining it clearly enough. Showing is easier than explaining, so this lowers the bar for automating the small, repeatable jobs that eat your week, like formatting a report or updating a spreadsheet the same way every Monday.

What to do about it: One real caution first, the recording captures your whole screen, so close anything private, like customer data or your banking tab, before you hit record.

China's Moonshot AI plans to publish the full open weights of its Kimi K3 model on Monday, July 27, a system that now sits at or above Claude and GPT on some coding leaderboards, according to VentureBeat. Open weights means anyone can download it, run it, and modify it for free, rather than renting access through a company. Rival DeepSeek is also due to ship a stable release of its V4 model this week at roughly $0.44 per million output tokens.

Why it matters to you: You will probably never download one of these yourself, and that is fine. What matters is the direction: frontier-level AI is becoming free and ownable, which drags down the price of every paid tool built on top of it. When the underlying engine is free, the companies charging you have to justify the bill.

Here is today's throughline: ads showing up inside ChatGPT, Google splitting its models to cut costs, Canva building websites from a sentence, Claude learning by watching, and a free frontier model landing Monday. Every one of these is the same move, AI pushing deeper into your daily work while the price of using it keeps falling. That is genuinely useful, and it is also exactly when the hype gets loudest and the "you must act now" emails start flooding in. You do not have to catch every car on this ride. Pick the one or two changes that clearly save you time or money, ignore the rest, and let the noise pass. Stay skeptical. Stay curious.

If this saved you from digging through five different AI newsletters just to figure out what actually matters, forward it to one person who needs to stop guessing about AI too. Someone forward this to you? Subscribe so the next issue lands straight in your inbox.

Talk tomorrow,
Mark Shilensky

Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social