Sponsored by
Two of the biggest AI labs on the planet just answered each other in public. Anthropic cut the price of its best model in half. Moonshot gave its best model away for free, no strings, no bill. If you are paying for an AI tool right now, this is the week to check whether you are still getting a fair deal.
Save 10+ Hours a Week With 37 Claude Prompts
Every manager faces the same situations before lunch: a message to land, a meeting to run, a hiring call, a report due. The AI Report built 37 Claude prompts for exactly those moments, organised by the situations every manager faces.
Copy the prompt, fill the brackets, run it in Claude, and get back 10+ hours a week. Oh, and it's free.
All you have to do is subscribe to The AI Report, a 5-minute daily AI brief read by 400,000+ business leaders at IBM, AWS and Microsoft, and the full prompt pack lands in your welcome email. The newsletter and the prompts, both free. Subscribe and grab both
Anthropic Cuts Its Smartest AI's Price in Half
What happened: Anthropic released Claude Opus 5 on July 24, priced at $5 per million input tokens and $25 per million output tokens, the same rate as the model it replaces. It lands within half a percent of Anthropic's own top-tier Fable 5 model on coding benchmarks, at half the cost, and it is now the default model on Claude Max and the strongest option on Claude Pro. A new "effort" setting lets you dial the model between fast and cheap or slow and thorough.
Why it matters to you: If you or your team use Claude for anything, from drafting emails to building simple automations, you likely just got a meaningfully smarter tool at no extra cost. That is not how software pricing usually works. Enjoy it while the price war lasts.
What to do about it: If you were avoiding Claude's top-tier model because of cost, check your account settings this week. You may already have access to more capability than you are using.
A Judge Just Priced Every Pirated Book Anthropic Used to Train Claude
What happened: A federal judge granted final approval of Anthropic's $1.5 billion settlement with authors and publishers, closing the largest copyright class action in U.S. history. Anthropic will pay $3,000 for each of an estimated 500,000 books it stored from pirated sources while training Claude. A separate 2025 ruling had already found that training on lawfully acquired books counts as fair use, but storing pirated copies in bulk does not.
Why it matters to you: This is the first real price tag the courts have put on unlicensed training data, and it is not a rounding error. If you use AI tools that were trained on scraped content, this case is the template other lawsuits will follow. The bill for cutting corners on data just became a real, calculable number instead of a hypothetical.
China Just Gave Away the World's Largest AI Model, and It's Free
What happened: Moonshot AI released free, downloadable weights for Kimi K3 on July 26, a day ahead of schedule. At 2.8 trillion parameters, it is the largest open-weight AI model ever made publicly available, and it already outranks Anthropic's Fable 5 on several coding leaderboards. Anyone can download it, modify it, and run it on their own servers for free.
Why it matters to you: Free and frontier-level used to be mutually exclusive. They are not anymore. This does not mean you should rip out your current AI tools, most business owners do not want to run their own servers, but it does mean the AI tools you pay for need to justify that price against a genuinely capable free alternative.
Nvidia Rallied 25 Companies to Defend Free AI Models. Anthropic Would Not Sign.
What happened: Nvidia, Microsoft, Meta, and 22 other companies signed an open letter urging Washington not to place broad restrictions on open-weight AI models. The letter landed days after the White House accused Moonshot AI of building Kimi K3 by copying outputs from Anthropic's own model, a technique called distillation. Anthropic, whose model was allegedly the one copied, was the notable holdout that refused to sign.
Why it matters to you: This is a fight over the rules for AI, not just the models. If the government restricts open-weight models to punish copying, it also restricts the free and low-cost tools many small businesses have started to rely on. Watch this one, the outcome will shape what AI options are legally available to you next year.
An Image Generator Just Learned to Control Robots
What happened: Black Forest Labs launched FLUX 3, a model trained jointly on images, video, audio, and robot actions inside one system. It can generate 20-second videos with synchronized sound from a text prompt, and the same underlying model is already being tested by Audi to help robots handle flexible parts like cables and seals on a production line.
Why it matters to you: The line between "creative AI tool" and "AI that runs physical equipment" is getting blurry fast. If your business touches manufacturing, logistics, or anything with moving parts, the tools built for marketing content today are the same architecture showing up on factory floors next.
A Defense Startup Wants to Be Worth More Than Northrop Grumman
What happened: Anduril Industries is reportedly in talks to raise funding at close to a $100 billion valuation, up from $61 billion just two months ago. That would put the AI-powered defense company ahead of Northrop Grumman, an 87-year-old defense contractor.
Why it matters to you: Money is not just flowing into chatbots. Investors are betting some of the biggest AI returns will come from physical, high-stakes industries like defense. That is a signal worth noting if you have written off AI as a "software only" trend in your own industry.
Sam Altman Says We're Living in the Singularity. Take That With a Grain of Salt.
What happened: OpenAI's Sam Altman told the "Relentless" podcast that "we are now, like, in the singularity," pointing to AI systems that recently helped crack decades-old math and physics problems.
Why it matters to you: Big claims from AI lab CEOs tend to arrive right when those same labs need headlines, funding, or both. The underlying research gains are real. The framing is marketing. Judge the tools by what they actually do for your business, not by how big the word attached to them is.
The Bottom Line
Every story this week points at the same fork in the road. One path gets cheaper and more open. The other gets tied up in courtrooms, policy fights, and CEOs reaching for bigger words to describe what they built. Neither path requires you to panic or to buy anything new today. It just requires you to keep asking what you are actually paying for, and whether a free or half-price alternative already does the job. That question will save you more money this year than any new tool will make you.
Enjoying the Ride?
If this issue was useful, forward it to one business owner who is still guessing about AI instead of reading about it. New here? Subscribe and get tomorrow's issue straight to 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