Hello, and welcome to TechScape. I’m your host, Blake Montgomery, The Guardian’s US tech editor, writing to you while cheering on Team USA in the Winter Olympics.
Introducing our new series about AI and the future of work
Throughout 2026, The Guardian will publish a series of stories about how artificial intelligence is affecting modern labor. We’re calling it Reworked: A series about what’s at stake as AI disrupts our jobs.
Our first story published this morning. Artificial intelligence, in particular the increasing automation of coding, has changed the attitude in the tech industry from one of relentless optimism and quirky perks like office ball pits to a default of grinding and austerity. Arielle Pardes reports on San Francisco’s new work ethic:
In the last year, as the magic dust of artificial intelligence has settled in the City by the Bay, the vibe among tech workers does seem different. The excitement about a new epoch in tech – and all the money that comes with it – is now tempered with anxieties about the industry and the economy. Some workers are going all-in on AI while also questioning whether all that AI is good for the world. Others are effectively training machines to do their jobs better than they can. And many of the same workers who are racing to build the future are now wondering if the future they’re building has a place for them in it.
The work of Mike Robbins, an executive coach in Silicon Valley who has worked with companies like Google, Salesforce, and Airbnb, illustrates the change. Robbins used to be asked to speak to companies and their leaders about topics like employee burnout, wellbeing and belonging – top priorities in the years during and shortly after the pandemic.
“Quite frankly, we’ve stopped talking about all that,” he says. Now, company leaders want advice on topics like change, disruption and uncertainty in the workplace.
Rather than a model of how we should all work, the tech industry may be a premonition for the anxiety and attempts to compensate that are coming for all of us. The change afoot in San Francisco serves as an early alarm to other industries – my own included – as AI encroaches on all types of work. Automating work rarely opens time for leisure. Instead, it increases expectations of productivity. Sometimes those goals rise to unattainable levels if an employer believes the technology is more powerful than it really is. Bosses’ unrealistic expectations are the subject of a story coming later this week in the series.
Stay tuned for more excellent pieces as Reworked launches: An essay about the bunk promise of the four-day work week by longtime Guardian columnist Robert Reich; on-the-ground tales from UK workers who have been training their own robotic replacements; and a powerful essay by the series’ editor, Samantha Oltman, about worker power in the age of artificial intelligence. You won’t want to miss them. You’ll be able to find them all on the series landing page.
This week in Elon Musk (and his brother Kimbal)
The AI industry rolls over speedbumps
Mandatory Credit: Photo by Idrees Abbas/SOPA Images/REX/Shutterstock (13657934j)In this photo illustration, a logo of samsung mobile company is seen on a mobile phone screen.Smartphone companies illustrations in India – 8 Dec 2022 Photograph: Idrees Abbas/SOPA Images/REX/Shutterstock
The high-velocity artificial intelligence sector hit two speedbumps last week. Each is quite different from the other, but both seem likely to be solved the same way.
The first snag: A memory chip shortage. Long, long ago (September 2025), experts predicted that data centers’ rising demand for computer memory would make your next smartphone more expensive if Donald Trump’s tariffs didn’t get there first. That prophecy has come to pass. The three most prominent memory chip producers–Samsung and SK Hynix of South Korea and Micron of Idaho, USA–have declared code red, Morning Brew reports. They can’t meet demand, which is growing because of the rapid rollout of AI infrastructure across the world. The price of consumer electronics is already increasing as a result, since computer memory chips are foundational to nearly every type of advanced device. Sony may debut the next Playstation due to the shortage. An observation: Data centers seem to me to be constructed not so much out of steel and concrete but out of pure hunger and thirst, buildings of reified need for money, power, water and every kind of computer chip–all of it never enough.
The second bump for the industry was a wave of departures. At Elon Musk’s xAI, multiple co-founders departed amid the reorganization and absorption of the company by SpaceX. Elsewhere, a leading safety researcher quit Anthropic, which makes the Claude chatbot and coder, to become a poet. OpenAI fired a safety researcher who opposed the introducing erotic content into ChatGPT’s responses for alleged sex discrimination. Another OpenAI researcher quit over the company’s decision to insert ads into ChatGPT.
It seems that the AI industry will overcome the shortages of memory and executives by the same means: money. Last week, I wrote about tech giants’ plans to spend some $600bn in the coming year alone. That amount of money exerts a gravitational pull like a black hole. Memory chip makers will sell their wares to the highest payer. Skilled employees will do the same. The losers in the memory chip shortage will be everyday consumers who need to replace their phones.
As for the executives, Musk seems to have had little trouble attracting talent despite his fearsome and multi-headed public controversies. At OpenAI and Anthropic, we have seen the departure of safety-minded executives result in little but their replacement and the continued debuts of products with greater power and potential for misuse. There is money to be made, after all, billions and billions, and this is the United States. The gravity of the cash grab in AI has repeatedly trumped ethical concerns and will do so again.
There does appear to be a looming crisis in AI, though, one which may reveal the true character of a firm that has cast itself as one of the most conscientious in the industry. A concrete version of the choice between safeguards and violent action lies before Anthropic now. The US military deployed Claude in its raid on Venezuela that resulted in the capture of Nicolás Maduro, according to the Wall Street Journal. That use doesn’t seem to have been satisfactory, though: The Pentagon is now considering cutting ties with Anthropic over the safeguards the startup has placed on its AI, Axios reports. After months of difficult negotiations between the Pentagon and Anthropic leadership, the startup has maintained two red lines it will not cross: the use of its technology for the mass surveillance of Americans or fully autonomous weaponry. The US military is getting fed up. Which side will balk?
Anthropic was spun out of OpenAI by safety-conscious executives and marketed to the public as a more mindful version of the ChatGPT maker and its ruthless founder. Will the young startup choose money or its corporate morals? Google faced a similar dilemma years ago with Project Maven, in which the US military used open source Google software to better identify people during drone flights. The tech giant dropped the project but later dropped its commitment not to use AI for weapons.
