Saturdays-are-for-Claude-How-AI-limits-are-quietly-reshaping.jpeg

Saturdays are for Claude: How AI limits are quietly reshaping the workday

Max Johnson used to start his workdays by opening a Claude chat and staying in it for hours.

The cofounder of Briix, a UK startup that helps small businesses use AI, would move fluidly from writing scripts for social media posts to designing graphics to generating documents — all in one long-running thread.

That lengthy dialogue meant he burned through more tokens, which are the currency of AI use, but giving so much context also sharpened the model. Johnson and his two cofounders could — and did — work “all hours of the day with no issues,” he told Business Insider.

That rhythm has changed.

In recent weeks, Claude’s usage limits have become more restrictive, to the point where Johnson, 24, said he can sometimes exhaust the allowance under his subscription “two prompts in” to a fresh chat. Now, he tries to sketch out his work around what can feel like an invisible meter.

“You plan your day around knowing that you can spend X amount of time,” he said.

Especially for users who subscribe to AI plans priced below enterprise accounts, the limits can be a challenge. AI companies, responding to the costs of running these models, are adjusting pricing.

In late March, Anthropic adjusted its usage caps during peak hours to manage demand.

An Anthropic spokesperson said in a statement that the company is adjusting its five-hour session limits to manage growing demand for Claude and has introduced efficiency improvements to offset the impact. About 7% of users will hit session limits they wouldn’t have previously. The company is also investing in capacity.

From entrepreneurs to software developers, usage limits on AI tools are reshaping how some workers structure their time, prioritize tasks, and think about work itself.

Workdays built around limits

For Johnson, having to wait, sometimes until an evening reset, can introduce a new kind of fragmentation in his day. At times, Johnson said, he and his cofounders will each hit their caps and then try to figure out what to do next.

“Panic sets in,” he said. They might take 30 minutes to an hour to consider how to move ahead.

The team might grab food, though the timeout isn’t necessarily welcome.

“It still doesn’t feel like a break because all I’m thinking about is, ‘When is this limit going to reset and we can get back to it?'” Johnson said.

Instead of relying on long, memory-rich chats, Johnson now often breaks things into smaller projects. For his social media work, there might be one for scriptwriting, another for animations, another for documents, each with tightly scoped instructions to conserve tokens.

Johnson, who is the company’s only full-time worker, previously shared a single Claude subscription plan with his cofounders. It made it easy to see what others were working on. Now, they’ve moved to individual accounts and, eventually, the five-person company in Manchester, England, might trade up to an enterprise plan, which he estimates could cost about $2,400 a year.

Saturdays are for the coders

For some developers, limits become part of the strategy.

Ani Potts, a 21-year-old New York University math major building a startup in stealth mode, treats his AI use like a weekly budget. He tries to plan around when he expects his allowance will reset by concentrating his work in high-intensity blocks.

After a day of classes, Potts often works in four-hour chunks. He saves the most demanding tasks — like research, testing, and any coding he might still do — for when he’s far from reaching his usage limit. As Potts nears the cap, he stops altogether or downshifts to small-ball concerns, like figuring out why a button on an app he’s developing is “a bit more blue” than he wanted.

Being pushed into an AI hiatus, Potts told Business Insider, can be “like going in slow motion,” yet he tries to think of it as a blessing.

“I can use my brain again,” Potts quipped.

The result is a rhythm where hitting the limit is expected — and sometimes useful. If Potts has a free Friday, for example, he’ll force himself to review his work and rethink priorities.

Saturdays, when he doesn’t have class, are for locking in. Rather than going to the bar or club, he said, “I Claude Code.”

Stopping work, by choice

When Danial Qureshi, a software developer in Toronto, hits a cap on his Claude Pro subscription, which costs him about 28 Canadian dollars a month, he often simply stops working on his personal projects.

“It’s basically not even worth my time to be manually writing code when I can have something like Claude doing it for me,” he told Business Insider. That’s because the AI output might be 10 times what he can do, he said.

Still, Qureshi, 27, sees an upside in the limit: By compressing what used to take hours of effort into short bursts of AI-assisted output, he feels sharper.

“Now I’m able to get more work done without feeling that cognitive burnout at the end of the day,” Qureshi said.

On weekends, he might spend a few hours building a project — like an AI agent that analyzes his jogging data and adjusts his training schedule — and then stop as he nears his limit for a five-hour window.

That means the rest of the day is open.

“You can actually go to the gym, meet friends, go to dinner, and then you still expend all the tokens, and you reach the usage limit,” Qureshi said.

Resetting expectations

For Johnson, AI tools have allowed him to dramatically increase what he expects to accomplish in a day. When he can’t do that, he said, it forces him to rethink his approach.

While Johnson still relies heavily on Claude and sometimes turns to ChatGPT as a backup, he expects he’ll still keep paying for Claude.

In the meantime, AI pauses could continue to punctuate Johnson’s workday.

“Let’s have some food now,” he said. “Let’s chill — wait for the limit to reset — and then we’ll get back to work.”

Do you have a story to share about session limits? Contact this reporter at tparadis@businessinsider.com.




Source link

Image of Lakshmi Varanasi

Claude hits No. 1 on App Store as ChatGPT users defect in show of support for Anthropic’s Pentagon stance

While OpenAI locks down Washington, Anthropic is locking down users and rocketing to the top of the App Store.

Anthropic has been sidelined in Washington following a public dispute with the Department of Defense over how its AI models would be deployed. President Donald Trump ordered federal agencies to phase out its technology.

Meanwhile, OpenAI has secured new ground, with CEO Sam Altman announcing in a Friday night post on X that it had reached an agreement with the Department of War to deploy AI models in its classified network.

OpenAI’s agreement has left some loyal ChatGPT users uneasy about OpenAI’s ambitions, prompting online debates about the ethical implications — and some saying they were defecting to its rival Claude.

As of 6:38 p.m. ET on Saturday, Claude ranked number one among the most downloaded productivity apps on Apple’s App Store, trailing ChatGPT.


A screencap of the app store

BI



Converts have taken to social media to share screenshots documenting their switch.

Pop musician Katy Perry wrote that she was “done” on X, alongside a screenshot of Claude’s pricing page, with a red heart around the $20-per-month “Pro” plan.

Another X user, Adam Lyttle, wrote “Made the switch,” alongside a screenshot of his email inbox with a receipt from Anthropic and cancellation confirmation from OpenAI.

On Reddit’s ChatGPT subreddit, dozens of users say they’ve deleted their accounts and are urging others to do the same.

“Cancel ChatGPT” has become a common refrain online, while some users have taken a more personal tone, saying Altman’s move “crossed the line.”

The agreement hasn’t polarized all AI users, however.

In one Reddit thread, several commenters said the news does not affect their choice of AI model, arguing that Anthropic’s work with Palantir raises similar concerns. In November 2024, Anthropic, Palantir, and Amazon Web Services struck an agreement to provide US intelligence and defense agencies access to Claude models.

After Secretary of War Pete Hegseth said he would designate Anthropic as a “supply chain risk to national security,” Anthropic said it would “challenge any supply chain risk designation in court.”

In his Friday post, Altman said the Department of War had agreed with two of OpenAI’s safety principles.

“Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems,” Altman wrote on X. “The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement.”

By Saturday afternoon, OpenAI published a more detailed description of its contract with the DoW, including the specific language it used surrounding the use of its models for surveillance and autonomous weapons.

On the topic of autonomous weapons, OpenAI said:

The AI System will not be used to independently direct autonomous weapons in any case where law, regulation, or Department policy requires human control, nor will it be used to assume other high-stakes decisions that require approval by a human decisionmaker under the same authorities.

On the topic of mass surveillance, OpenAI said:

The AI System shall not be used for unconstrained monitoring of U.S. persons’ private information as consistent with these authorities.

While some chatbot users suggested it’s all fair in business, war, and federal procurement, others suggested the Pentagon’s stance may have handed Anthropic a public relations win.

X user Tae Kim joked that Hegseth might need a new title: “Secretary Hegseth Chief of Claude Marketing.”




Source link

Anthropic-says-its-buzzy-new-Claude-Cowork-tool-was-mostly.jpeg

Anthropic says its buzzy new Claude Cowork tool was mostly built by AI — in less than 2 weeks

Anthropic’s new working agent was largely built by Claude itself — the latest example of AI coding tools speeding up product development.

On Monday, Anthropic announced the release of Cowork, a “more approachable” AI tool accompanying Claude Code that’s geared toward fulfilling users’ requests that are unrelated to programming. Users grant the agentic AI tool access to specific files on their computer and prompt it to complete tasks.

Boris Cherny, head of Claude Code, said that Anthropic’s AI coded “pretty much all” of Cowork.

“@claudeai wrote Cowork,” Product Manager Felix Rieseberg wrote on X. “Us humans meet in-person to discuss foundational architectural and product decisions, but all of us devs manage anywhere between 3 to 8 Claude instances implementing features, fixing bugs, or researching potential solutions.”

As a result, Rieseberg said the first edition of Cowork came together quickly.

“This is the product that my team has built here, we sprinted at this for the last week and a half,” he said during a livestream with Dan Shipper.

Over the holidays, Rieseberg said that Anthropic saw its customers using Claude for an increasing number of non-coding-related tasks.

“This sort of like the research preview, very early Alpha, a lot of rough edges, as you’ve already seen, right?” he said.

Cowork is initially available to Claude Max subscribers on the Mac app.

The launch has made a splash in the tech world, with many online users praising the product and its accessibility.

“I think that’s a really smart product,” Datasette co-creator Simon Willison wrote in a blog about his experience. “Claude Code has an enormous amount of value that hasn’t yet been unlocked for a general audience, and this seems like a pragmatic approach.”

“This is big,” Reddit cofounder Alexis Ohanian wrote on X.

Because granting an AI agent access and the ability to take action on specific computer files comes with risk, Anthropic cautions that Cowork users should be careful.

“By default, the main thing to know is that Claude can take potentially destructive actions (such as deleting local files) if it’s instructed to,” the company said. “Since there’s always some chance that Claude might misinterpret your instructions, you should give Claude very clear guidance around things like this. “

The latest in a flurry of AI announcements

AI companies wasted no time in launching new offerings and partnerships to kick off the new year.

On Sunday, Anthropic announced Claude for Healthcare, a major addition to its healthcare and life sciences offerings. Its release came on the heels of rival OpenAI signaling its investment in the healthcare space with ChatGPT Health.

Amid AI bubble chatter and scrutiny on the increasing AI investments made by tech companies, Anthropic CEO Dario Amodei has argued that Anthropic has built a more sustainable business model that allowed it to make more educated bets on its future build-out. While he did not name OpenAI or CEO Sam Altman directly, he made some thinly veiled criticisms of his former company throughout the event.

“I think because we focus on enterprise, I think we have a better business model,” Amodei said at The New York Times’ Dealbook Summit. “I think we have better margins. I think we’re being responsible about it.”

Google, which some experts saw as overtaking OpenAI at the end of 2025, announced a major deal with Apple to have Gemini power Siri’s artificial intelligence capabilities.




Source link

Henry Chandonnet is pictured

The creator of Anthropic’s Claude Code likes to hire engineers who do ‘side quests’ like making kombucha

Want a job at Anthropic? It might help to get a hobby.

The AI boom is changing the job requirements for an engineer. Not only do they need to have coding skills, but they also must know how to operate vibecoding tools and stay up to date with new AI models.

Anthropic leader Boris Cherny looks for something else: “Side quests.”

“When I hire engineers, this is definitely something I look for,” he said on “The Peterman Pod.”

Cherny’s definition of side quests includes “cool weekend projects,” like someone who’s “really into making kombucha.” It’s a sign that the engineer is curious and interested in other things, he said.

Much of Cherny’s own growth came from his side projects. Cherny is now a key figure at Anthropic. He created Claude Code, a tool that is now popular with engineers across the country.

“These are well-rounded people,” he said. “These are the kind of people I enjoy working with.”

Cherny also said he prefers that his new hires be “generalists.”

He gave the example of an engineer who can code, but is also able to work on product and design. That all-star engineer also seeks out user feedback.

“This is how we recruit for all functions, now,” he said. “Our project managers code, our data scientists code, our user researcher codes a little bit.”

Cherny isn’t alone in pushing for jobs to become more generalist. Figma CEO Dylan Field said in October that AI was causing job titles to merge, resulting in everyone being a “product builder.”

What else is Anthropic looking for? For some time, it monitored whether candidates use AI in their applications.

In May, Business Insider reported that Anthropic asked candidates for certain jobs not to use AI in their written responses so the company could test their “non-AI-assisted communication skills.”

Anthropic changed its policy in July, allowing candidates to seek out assistance from Claude.

For the younger engineers, a job at Anthropic may be hard to come by. In May, CPO Mike Krieger said on “Hard Fork” that he was focused on hiring experienced engineers — and had “some hesitancy” with entry-level workers.

On the podcast, Cherny said that his love of generalists came from his career trajectory. Working at startups since 18, Cherny had to do everything, he said.

“At big companies, you get forced into this particular swim lane,” he said. “It’s just so artificial.”




Source link