Bill Gates Wants AI Taxed Before It Takes Your Job

Bill Gates spent decades building the machine that is now coming for other people's jobs, and this week he said the quiet part out loud. Today's issue also covers Salesforce putting Claude inside your CRM, Apple building a Mac that runs a huge AI model with zero internet connection, and South Korea handing every citizen free AI. Let's get into it.

Creators lock in their holiday content calendars 90 days out, before most ecommerce brands finalize their commission strategy and way before Black Friday and October deal events.

Get ahead of the seasonal rush with The 90-Day Holiday Sprint, a practical guide for brands that want creators driving holiday demand while competitors are still recruiting:

Your 90-day countdown starts now.

Get the Free Guide

What happened: Bill Gates published a new essay this week calling the AI era "one of the most turbulent times in human history," arguing that unlike the shift to computers or the internet, this one is moving too fast for people to adapt (Tom's Hardware). He named three risks: job losses hitting workers who cannot easily retrain, AI making scams and cyberattacks easier, and heavy AI use weakening kids' social skills. His proposed fix: designate some jobs "Human Reserved," similar to a nature preserve, and tax AI tokens and robots to replace the income tax revenue lost as machines take over paid work.

Why it matters to you: If you run a small business and lean on people for skilled, judgment-heavy work (a mechanic, a bookkeeper, a caregiver, a tradesperson), Gates is telling governments those are exactly the roles worth protecting on purpose, not by accident. It also means the tax and regulatory ground under AI use could shift faster than you'd expect, especially if you're automating a role instead of adding a tool to it.

What to do about it: If you're weighing whether to replace a role with AI instead of adding AI to a role, that's the same line regulators are starting to draw. Keep it in mind before you make the swap permanent.

What happened: Anthropic and the Janelia research institute opened a research preview of the Model Hardware Standard, a shared way for AI agents to plug into and operate physical lab equipment instead of just software (Anthropic). Early testers saw real results: a Carnegie Mellon lab ran experiments 3 times faster, and at QuEra, an AI agent fixing a broken quantum laser lock pushed the recovery success rate from 58% to 99.3%, cutting the fix time from 150 seconds down to about 6.

Why it matters to you: This isn't really about lasers. It's a preview of AI agents doing physical, hands-on troubleshooting, not just answering questions on a screen. If your business runs equipment or machinery with a recurring "it broke again" problem, this is the direction things are heading, even if it's years from your shop floor.

What happened: Salesforce and Anthropic launched "Claudeforce," putting Claude directly inside Salesforce with 37 ready-made sales skills, including pipeline reviews, deal health checks, and meeting prep, pulled from your actual CRM data instead of generic chatbot guesses (Salesforce). It opens in beta in September, with one company-wide setup instead of configuring it person by person.

Why it matters to you: If you or your sales team already live in Salesforce, this could save real hours on the grunt work of prepping for calls and chasing pipeline updates. If you don't use Salesforce, it's still worth watching, because every major CRM is racing to bolt AI onto the same jobs.

What to do about it: If you're on Salesforce, ask your admin when the beta opens for your account, so someone tries it before you pay for a separate tool that does the same thing.

What happened: Anthropic made Claude in Chrome available to everyone on a paid Claude plan (Claude). It can read a webpage, click links, fill out forms, and move between tabs using your existing logins, and it now handles multi-step tasks on its own instead of stopping for approval at every click, with a safety check reviewing each action first.

Why it matters to you: A lot of the tedious, repetitive work business owners do online (filling out the same form on five directories, pulling numbers off a dozen webpages, updating listings) is exactly what this is built for. It's a real shortcut, not a novelty, but it's Chrome-only for now, so it won't replace your phone workflow.

What to do about it: If you already pay for Claude, install the Chrome extension and try it on one repetitive task this week before deciding it's worth building a habit around.

What happened: South Korea picked SK Telecom, KT, and Kakao to run a new "AI for All" program that gives residents free, unlimited access to AI tools for things like scheduling doctor visits, home searches, taxes, and general help (Korea Times). The government is shipping 512 Nvidia B200 chips to power it this year, with a full public launch planned for late 2026 after a beta period, and it's spending about $7.2 billion on AI this year, three times what it spent last year.

Why it matters to you: This isn't your market, but it's a preview of how fast a government can move once it decides AI access is basic infrastructure, like electricity or water. Watch for versions of this argument to show up in U.S. state and local policy debates, especially anywhere officials are deciding who gets AI tools and who doesn't.

What happened: OpenAI began showing ads inside ChatGPT's free and low-cost "Go" tier in India, working with roughly 50 brands through ad agencies WPP and Omnicom, with a self-serve ad manager for any advertiser launching in September (TechCrunch). OpenAI already rolled ads out to U.S. users earlier this year and expanded to Europe in August, so India is the next stop, not the last one.

Why it matters to you: If you've been using free or cheap ChatGPT tiers for your business, expect ads to show up in your own chats eventually, the same way they did on free email and social apps. Once the self-serve ad manager opens more broadly, small businesses may also get a new place to advertise directly inside AI chat itself.

What to do about it: If ads inside your AI tools would bother you, weigh that against the cost of a paid tier now, before ads become the default everywhere.

What happened: Apple introduced new M6 and M5 Ultra chips, with the top Mac Studio configuration now packing up to 512GB of shared memory (Apple). That's enough to run huge AI models, the kind that normally require renting space in someone else's data center, entirely on a machine in your office, with no cloud subscription and no internet connection required.

Why it matters to you: For any business worried about sending customer data to a cloud AI service, or just tired of stacking monthly AI subscription fees, this is a real alternative on the horizon: buy the hardware once, run your AI locally, and keep your data in the building. It's expensive hardware, but the math might work if you're already paying for several AI subscriptions.

What happened: Producer Dr. Dre confirmed he's using AI in the studio and said only people "who have trouble creating" see it as a threat, comparing the backlash to the panic over drum machines decades ago (Variety). Music industry veteran Jimmy Iovine added that plenty of artists are already quietly using AI tools without admitting it publicly.

Why it matters to you: The loudest voices against a new tool are often the ones who feel most threatened by it, not the ones with the clearest read on where it's actually useful. If you've been avoiding AI tools in your own business because it feels like cheating, it might be worth asking whether that hesitation is really about quality, or just about getting comfortable with something new.

None of today's stories are really about AI getting smarter. They're about AI getting installed: into your CRM, into your browser, into a government's basic services, into a Grammy winner's studio. The rollercoaster isn't the technology anymore. It's how fast the ground underneath your business is moving while you're still deciding whether to strap in. You don't have to chase every drop, but you do have to know which ones are actually headed your way.

If this issue was useful, forward it to one person who needs to stop guessing about AI and start understanding it. New here? You can get your own copy free tomorrow by subscribing.

Talk tomorrow,
Mark Shilensky

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