Price beat loyalty, two models picked parties, and a government picked a contract over a chatbot. Today rewards whoever reads the fine print.
Today's five drop in 02:34:53
Rippling now sends 52% of its engineers' coding work to OpenAI models, up from 16% in May, as OpenAI discounts its way into Anthropic accounts, The Information reports. The price war has customers now, and you can use it.
Why it matters If you spend real money on Claude or GPT, the list price is now an opening bid. Get a written quote from a second lab before your next renewal, and keep your stack able to switch models in a day, not a quarter.
Asked through the API who to vote for in 479 midterm races, GPT-6.1 Sol named the Democrat in 97% of its picks and Grok 4.7 the Republican in 92%, a Convovo study found. Your model has a ballot, whether you asked for one or not.
Why it matters If you build anything that answers questions about candidates, the model you choose brings its own lean, and a "just decide" prompt exposes it. Test your product on the races your users will ask about before Nov. 3, and decide what it should refuse.
SRCConvovo
Sweden's Government Offices will block ChatGPT and Claude on staff computers starting Oct. 26 and steer civil servants to Microsoft Copilot, Decrypt reports, citing Sveriges Radio. The ban is about contracts, not capability.
Why it matters If you sell AI to a government or any regulated buyer, the model is not what gets you in the door. The signed data agreement is. If you use a consumer chatbot at work without one, expect your IT team to follow Stockholm.
SRCDecrypt
Apollo chief economist Torsten Slok says revenue per worker is rising for the Magnificent 7 but flat for the rest of the S&P 500, with no sign yet of AI lifting margins or productivity elsewhere. The spending is real. The proof has to come from companies like yours.
Why it matters If you run or defend an AI budget, the burden of proof just moved to you. Pick one workflow, measure its output in hours or dollars before and after AI, and have the number ready before your CFO reads this chart.
SRCApollo
Meta pushed its new Commercial Terms, which let it crawl websites linked to business accounts, from Oct. 30 to Nov. 20 after publishers objected. It already dropped the word "train," but the crawling clause is still there, so read it before then.
Why it matters If your business has a Facebook or Instagram account, these terms reach the website you link to it. You have until Nov. 20 to read the clause, check your robots.txt, and decide whether you are fine with Meta's crawlers on your pages.
SRCMeta
Build a plan to cut your AI bill or switch providers
› I want to negotiate a better price on my AI model spending, or switch providers if that turns out cheaper for the same quality. Help me build a one-page plan.
My situation:
- Current provider and the models I use most: [provider, models]
- Monthly spend: [$ amount]
- Main workloads: [e.g. coding agent, support replies, document summaries]
- Contract or plan renewal date: [date]
Do this:
1. Rank my workloads from least to most risky to move to another model, with one line on why for each.
2. Write a 10-task test set for my main workload that I can run on two providers side by side, plus a simple 1 to 5 scoring rubric.
3. Draft a short, polite email to my current provider asking for a better rate, saying I'm evaluating alternatives. No bluffing about offers I don't have.
4. List the switching costs I'm likely forgetting: prompt rewrites, re-running tests, rate limits, data and privacy terms, team retraining.
5. Finish with a decision rule: at what price difference and what quality gap should I switch, and when should I stay?
Ask me for any missing detail before you start.Works in any AI chat. In Magai, run it against several models at once.
Each square is an edition. Click any day to read the five that mattered then.
Every term our stories use, defined in plain English and linked wherever it appears.
The glossary →