AI + Careers Briefing — November 5, 2026

Today's three stories show AI at work running into budgets and real behavior. Apple is reportedly putting a daily ceiling on what staff can spend on AI. Uber says almost no customers use AI agents to book a ride. And the tools for AI-written code are being bought up by a handful of big platforms. For job seekers, the common thread is that results you can measure, and skills that carry across tools, count for more than how much AI you use.

Is Apple really cutting how much its employees can spend on AI?

Apple has reportedly lowered the daily amount its engineering teams can spend on AI tools, according to a Bloomberg report that TechCrunch summarized on Nov. 4.

TechCrunch wrote that, per Bloomberg, Apple cut the daily AI spending cap for engineering teams in October from $300 to $200. Bloomberg's sources said further cuts are planned, with some allowances dropping to $100 a day. Outside engineering, Apple is reportedly considering a daily limit as low as $25.

TechCrunch said it contacted Apple for comment, and its article does not include a response. We could not access the Bloomberg original, so every Apple figure here comes through TechCrunch's account and remains unconfirmed by Apple.

TechCrunch framed the move as a shift for a company that has worked to build AI into its engineering processes. It also described a broader pattern: earlier in 2026, companies pushed staff to use AI as much as possible, a habit it called "tokenmaxxing," and some have since begun rationing tokens as the bills arrive.

Separate research points the same way on cost. HR Dive reported on Oct. 29 that Gartner's HR practice found 38% of companies are over budget on AI and 85% plan to spend even more next year, while Gartner's senior director analyst CV Viverito said organizations will not generate cost savings from AI at this point. That is Gartner's finding on companies in general, not on Apple.

Why it matters for job seekers: This section is our read, not reporting. If employers are putting daily ceilings on AI use, "I use AI constantly" becomes a weaker pitch than "I get measurable results from AI at a sensible cost." In interviews, we would lead with an outcome (time saved, errors caught, work shipped) and be ready to explain which tasks justify an AI tool and which do not. Candidates who can describe the result of their AI use, rather than the volume of it, look better suited to a workplace where usage may be capped. One caveat: Apple is a single reported case, and it does not show that other employers are following.

How Many People Actually Use AI Agents to Book an Uber?

Almost none: Uber said only four rides were ordered through an AI agent during its busiest week of October, the week of Halloween, according to TechCrunch's report on the company's third-quarter earnings call.

Food delivery fared slightly better. Uber CFO Balaji Krishnamurthy told analysts the company handled around 320 million trips and orders across Uber and Uber Eats that week, and that only 500 Uber Eats orders were placed by an AI agent. The 320 million figure combines rides and food orders, so it is not a rides-only denominator. A call transcript published by Investing.com has the same figures, with Krishnamurthy describing the traction so far as very early and very limited.

TechCrunch also reported that booking a ride in Uber's app takes about 50 seconds on average and reordering a favorite food takes about 90 seconds, and its reporter suggested that means agents may not save users much time. That is the reporter's inference, not a company statement.

The numbers land amid heavy investment in agent products. TechCrunch noted that OpenAI, Anthropic and Google offer AI agents, and that Meta said its Muse app reached the top of Apple's U.S. App Store last month. It also reported that DoorDash said on its earnings call that it received only a few hundred orders from AI agents, with most customers still ordering through its own app.

Why it matters for job seekers: This is our read, not Uber's claim. One company's week of data does not settle how fast agents will spread, and Uber's leadership remains optimistic about the long run. But the gap between agent buzz and agent use is a useful check on how you read AI claims in job postings and interviews. When an employer says agents will transform a role, ask what share of real customers or workflows use them today. A company that can give you numbers is telling you something different from one that can't. We also think the work most likely to stay in demand is the work that makes AI function inside real products, such as integration, evaluation and keeping systems reliable, because adoption depends on it. For ideas on which skills hold up, see our guide to AI-proof career skills.

Why did Meta and OpenAI both buy AI coding startups on the same day?

Meta and OpenAI each announced acquisitions of AI-coding companies on November 4, 2026: Meta confirmed it is buying Conductor, and OpenAI said it will acquire Pierre Computer Company to expand its Codex coding product.

TechCrunch reported that Conductor, founded in 2024 by CEO Charlie Holtz, lets users build software by running multiple AI agents in parallel workspaces. Those agents come from major labs, including Anthropic and OpenAI, rather than being built in-house. Meta did not disclose terms. Conductor's cloud service shuts down November 20, after which user cloud data will be deleted, and its core features keep working until early next year. TechCrunch also noted that Meta launched its own coding agent, Muse Code, in August 2026.

OpenAI's announcement describes Pierre as the company behind Code Storage, a service for storing code and tracking its history using Git, plus open-source developer tools. OpenAI said the Trees and Diffs libraries Codex already uses for browsing files and reviewing changes come from Pierre. It said Pierre's open-source projects will remain open-source and that customers can keep using Code Storage after closing. The page gives no price. It says closing is subject to customary conditions, including regulatory approvals, and that the companies remain separate until then.

Why it matters for job seekers: This paragraph is our read, not reporting. Both deals are about the tooling around AI-written code: Conductor coordinates agents, and Pierre's tools help people navigate and review code. Tool ownership is concentrating in a few big platforms, and a product you build habits around can change or shut down quickly, as Conductor's November 20 cloud shutdown shows. In interviews and on resumes, we would emphasize skills that carry across tools: writing clear specs, reviewing and testing code an agent produced, and managing several tasks at once. Name the tools you have used, but present them as examples of those skills. For a broader look at which abilities hold value, see our guide to AI-proof career skills.

What to watch

• Whether Apple's reported further cuts, to $100 a day for some engineering allowances, go ahead, and whether Apple confirms the figures.
• Whether other companies share agent-usage numbers on their earnings calls, after Uber's four rides and DoorDash's few hundred orders.
• Conductor's cloud shutdown on November 20, and OpenAI's Pierre deal, which still needs customary closing conditions, including regulatory approvals.