Economists at Stanford University have found the strongest evidence yet that artificial intelligence is starting to eliminate certain jobs. But the story isn’t that simple: While younger workers are being replaced by AI in some industries, more experienced workers are seeing new opportunities emerge.
Erik Brynjolfsson, a professor at Stanford University, Ruyu Chen, a research scientist, and Bharat Chandar, a postgraduate student, examined data from ADP, the largest payroll provider in the US, from late 2022, when ChatGPT debuted, to mid-2025.
The new study reveals a nuanced picture of AI’s impact on labor. While advances in artificial intelligence have often been accompanied by dire predictions about jobs being eliminated—there hasn’t been much data to back it up. Relative unemployment for young graduates, for instance, began dropping around 2009, well before the current AI wave. And areas that might seem vulnerable to AI, such as translation, have actually seen an increase in jobs in recent years.
“It’s always hard to know [what’s happening] if you’re only looking at a particular company or hearing anecdotes,” Brynjolfsson says. “So we wanted to look at it much more systematically.”
By combing through payroll data, the Stanford team found that AI’s impact has more to do with a worker’s experience and expertise than the type of work they do. More experienced employees in industries where generative AI is being adopted were insulated from job displacement, with opportunities either remaining flat or slightly growing. The finding backs up what some software developers previously told me about AI’s impact on their industry—namely that rote, repetitive work, like writing code to connect to an API, has become easier to automate. The Stanford study also indicates that AI is eliminating jobs but not lowering wages, at least so far.
The researchers considered potentially confounding factors including the Covid pandemic, the rise of remote work, and recent tech sector layoffs. They found that AI has an impact even when accounting for these factors.
Brynjolfsson says the study offers a lesson on how to maximize the benefits of AI across the economy. He has long suggested that the government could change the tax system so that it does not reward companies that replace labor with automation. He also suggests AI companies develop systems that prioritize human-machine collaboration.
Brynjolfsson and another Stanford scientist, Andrew Haupt, argued in a paper in June that AI companies should develop new “centaur” AI benchmarks that measure human-AI collaboration, to incentivize more focus on augmentation rather than automation. “I think there’s still a lot of tasks where humans and machines can outperform [AI on its own],” Brynjolfsson says.
Some experts believe that more collaboration between humans and AI could be a feature of the future labor market. Matt Beane, an associate professor at UC Santa Barbara who studies AI-driven automation, says he expects the AI boom to create demand for augmentable work—as managing the output of AI becomes increasingly important. “We’ll automate as much as we can,” Beane says. “But that doesn’t mean there won’t be a growing mountain of augmentable work left for humans.”
AI is advancing quickly though, and Brynjolfsson warns that the impact on younger workers could spread to those with more experience. “What we need to do is create a dashboard early-warning system to help us track this in real time,” he says. “This is a very consequential technology.”
Alongside its price-friendlyiPhone 17e and M4 iPad Air yesterday, Apple just announced a few updates to the MacBook Pro, MacBook Air, and its rarely-refreshed desktop display line.
The MacBook Air has now been updated to the latest M5 chip. It’s a fairly modest upgrade, but it brings it up to speed with Apple’s latest processor that debuted in the MacBook Pro last fall. There are no other major hardware changes—it now comes with 512 GB of starting storage with “faster SSD technology”—but you can still get the Air in either a 13- or 15-inch screen size.
This laptop also features Apple’s N1 wireless chip, which includes Wi-Fi 7 and Bluetooth 6 for the latest connectivity standards. It still comes with the standard 16 GB of RAM, and sadly, there’s a $100 price bump to account for the extra storage. It now starts at $1,099 for the 13-inch model and $1,299 for the 15-inch model. Apple says you can preorder it tomorrow, with sales kicking off on March 11.
More interestingly, Apple is expanding the M5 chip series with the M5 Pro and M5 Max, now available in the 14-inch and 16-inch MacBook Pro. Like previous generations of Apple silicon, the “Pro” and “Max” configurations add significantly improved multi-core CPU and graphics performance.
The new MacBook Air with M5.
Photograph: Courtesy of Apple
The M5 Pro and M5 Max can be configured with up to 18 CPU cores (12 performance cores and 6 “super” cores), up from 16 on the M4 Max. The M5 Pro can scale up to 20 GPU cores, while the M5 Max extends up to 40 GPU cores. Thanks to higher memory bandwidth, more efficient Neural Engine, and improved GPU architecture, Apple says the M5 Pro and M5 Max have “over 4X the peak CPU compute for AI” compared to the last generation and offer 20 percent better GPU performance.
The new MacBook Pros don’t include any other hardware changes; things have stayed largely the same since 2021—same port selection, Mini-LED display, speakers, and webcam. Even the claimed 24-hour battery life hasn’t changed from the M4 models, which came out in late 2024. Interestingly, as recently as last week, Bloomberg reported that Apple plans to launch a more significant update to the MacBook Pro later this fall, which will reportedly debut the M6 chip, an OLED touchscreen, and a thinner chassis.
Like the MacBook Air, all versions of the M5 Pro or M5 Max MacBook Pros come with twice the storage and a slightly higher starting price. Coming with 1 TB, the 14-inch M5 Pro now starts at $2,199, and the 16-inch model at $2,699. That’s $200 more than last year’s machines. Meanwhile, M5 Max prices start at $3,599.
“Increasingly, I’m coming back to running product and working with the vice president of tech on some artificial intelligence (AI) projects and getting very hands-on myself,” says Wolf & Badger CEO and co-founder George Graham.
“It’s intellectually challenging, stimulating and intriguing – and I want to learn more about it. I’m trying to get as much info as I can on what I consider to be the most interesting tech advancement of my professional work lifetime.”
Not the words of a head of engineering, CIO, or technology executive, but those of the CEO of the online marketplace, whose business world continues to be lit up by the opportunity to use AI across multiple operations.
And why not? In January 2026, Wolf & Badger released a performance update to mark 15 years of trading, reporting it had now surpassed $500m in cumulative sales since inception and achieved almost 40 million website visits in 2025 alone, while reinforcing its reputation as an ethical platform by securing B-Corp recertification.
Wolf & Badger partners with independent brands promising strong ethics, and effectively becomes their tech stack and online operations provider. It is the conduit for these brands to achieve the scale that few organisations of their size can achieve alone.
The business achieved annual sales of $100m (£75m) in 2024, with more than 2,000 brand partners now in place, helping the London-headquartered operation grow globally. And ongoing investment in “AI-driven discovery and on-site personalisation” is delivering a measurable impact, with the company talking about £3.2m of directly attributable incremental sales from recent AI initiatives.
“There’s tremendous opportunity to improve the efficiency and discovery on Wolf & Badger by better understanding our shoppers and our brands and the products they sell,” he says.
“There’s lots around AI on image recognition and product tagging to build out better information related to style or what event a product would be suitable for, and using that to surface more relevant products to the user at the right time – all with the end point of making life more exciting and creating inspirational shopping experiences.”
Are we all product managers now?
While the work on using AI to power the online experience is not uncommon in e-commerce today, Graham’s attitude as CEO of the marketplace is. He is a CEO getting his hands dirty with the tech, which is rare in retail.
“I have personally spent many hundreds of hours over the past three months getting my head around AI and the future of commerce – with agentic commerce in mind,” he explains.
I have personally spent many hundreds of hours over the past three months getting my head around AI and the future of commerce – with agentic commerce in mind George Graham, Wolf & Badger
“Claude Code has become my go-to app. I have built a fairly bespoke AI agent ‘chief of staff’ that is connected to my tools via MCPs [model context protocols] or APIs [application programming interfaces], with a bunch of bespoke skills and scripts that I have ended up building into that.”
Graham says the collective memory stack is getting more powerful by the day, and using it has improved his own working practices.
“I feel twice as efficient as I was six months ago. I have taken that time and – in the short term – continued reinvesting it in understanding AI.”
The AI assistant Graham has developed is being made accessible via Stack internally, and the wider team is getting set up on Claude Code themselves with access to their own version.
The CEO acknowledges he isn’t an engineer or coder, but as a teenager, he would make games in Basic (Beginner’s All-purpose Symbolic Instruction Code) and design websites with HTML. After studying business at university, he joined PwC as a strategy consultant. By the age of 23, he was starting Wolf & Badger and “had to figure out how to build a marketplace as there wasn’t really marketplace software out there”.
When development at Wolf & Badger was brought in-house – today, it has a vice-president of technology, around a dozen people in engineering and others in product management and design – Graham continued to play a part in building out features to support brand partners and customers.
“I have always found all that fascinating,” he explains.
“Over the years, I stepped away as we brought the experts in – but, increasingly, I’m coming back to running product and working with our vice-president of tech on some of the AI projects and getting very hands-on myself.”
Graham, who founded Wolf & Badger with his brother Henry in 2010, admits he doesn’t fully understand the finer nuances of coding and doesn’t have the experienced engineer’s eye. But with the new tech available, he suggests “anyone can be a product manager or software developer” now.
“I have been able to create prototypes – I have built things that assess brands coming on the platform and help the sustainability team with vetting,” he says.
The software his internal teams are now using is set up on GitHub, and built on Eversell MSL front-end, Superbase, and other apps. “Everything is hooked up in what I think is reasonably robust for internal use,” he adds.
He urges other leaders in retail and wider business not to be afraid, and to experiment with the tools now available. There’s a lot that can be built on just a small monthly tech subscription outlay, he notes.
The wider tech team at Wolf & Badger initially experimented with solutions such as Microsoft Copilot and then Cursor.
“Only in the past few months have our engineers found the quality is at a point where they can lean on it more to start actually writing code. We’re keeping clear of the vibe coding in key sensitive areas, of course, but we can experiment in lots of spaces.”
The tech exploration work Graham has taken on – and there is much more to come, he says – is to “ready ourselves for agentic commerce and make sure we’re ahead of the pack”.
Since the turn of the year, there have already been some noteworthy developments in agentic commerce that further underline it as a future direction of travel for e-commerce and, therefore, something e-commerce and retail leaders must better grasp an understanding of.
Agentic commerce and UCP
The new year started with JD Sports announcing it is enabling consumers to use AI platforms to search for and purchase products – all in a single click, without leaving the apps.
JD customers in the US can purchase directly through Copilot, and – in due course – this will be followed by the ability to do so via Google Gemini and ChatGPT. JD is leveraging the agentic commerce suite of tech players Commercetools and Stripe.
Jetan Chowk, JD’s chief technology and transformation officer, said the move was about meeting customers “where they are”. It came after OpenAI announced in September that US shoppers could buy from Etsy directly through ChatGPT.
It started with an “instant checkout” to support single-item purchases, but multi-item carts are now on their way to being a reality.
Then, at the January 2026 NRF Big Show in New York, where many from the retail technology community congregate every year, Google launched the Universal Commerce Protocol (UCP), an open standard for agentic commerce that aims to establish a common language for AI agents and systems to operate together across consumer surfaces, businesses and payment providers.
The work Stripe is doing with agentic commerce protocol and standardising the mechanism by which people can shop directly via the [AI] agents is super interesting George Graham, Wolf & Badger
This is a fast-moving space, but was co-developed with prominent retail industry players such as Shopify, Etsy, Wayfair, Target and Walmart, and endorsed by more than 20 others across the ecosystem, including payments companies Adyen, American Express, MasterCard and Visa.
Graham is in close conversations with Stripe and Google, attending their events and regularly tuning into their updates.
“The work Stripe is doing with agentic commerce protocol and standardising the mechanism by which people can shop directly via the agents is super interesting,” he says. “Google and Shopify UCP is a further move towards a standardisation of how this is going to work.”
Graham is confident there will be more consumer discovery conducted on Google’s AI-powered platforms, ChatGPT, Perplexity, and other similar spaces.
“We need to ensure we’re supporting the 2,000 brands we’re working with to appear in the right way on those channels and facilitate the tech that can support one- or zero-click checkout, where an agent has the ability to buy on a consumer’s behalf.”
He is confident that a platform such as Wolf & Badger can play a key role in the agentic space. Individual brands are typically going to struggle to really build out the right metadata and set up UCP to be recognised by the human in the loop or an AI agent.
Graham says: “If we can wrap together the best independent brands and collectively go to a shopping agent to ensure those brands appear in the right places, we’re well placed to capture some of that demand and drive it towards the individual brands we work with, rather than the resulting purchases ending up with the bigger homogenised brands in our space.”
He adds that Wolf & Badger’s presence harks back to the pre-digital days of boutique shopping in-store, but with the right technology investment and focus now, it can deliver this in a “scaled way” online and through its showrooms.
“Our editorial and marketing team still make the creative calls, but we’re able to drive it forward with some of these new bits of tech,” he says, adding that as Wolf & Badger extends its technological nous, it can enable its brands to focus on “the difficult part” of commerce – meaning the design and manufacturing of compelling garments and consumer products.
Rapidly evolving space
As for the immediate future at Wolf & Badger, the US expansion is a key focus – as are ventures across Europe and into the Middle East. An expanded brand partnerships function within the business is expected to support the onboarding of new designers from around the world.
But AI continues to be an area of significant exploration, with Graham confident that his experimentation and use of cost-effective tools are improving how the business operates.
“It’s a rapidly evolving space – everything is changing these days,” he says, adding that it’s getting increasingly difficult to understand what will come next due to the acceleration of technological capability.
“You just have to try to stay ahead,” he says. “We’re repositioning ourselves in making sure we are embracing AI in the way I think any forward-thinking growth company should, and recognising the power it can bring to enable us to do much more for our brands and shoppers.”
Vendor or supplier lock-in has been a longstanding topic of discussion, as far back as my first days in IT all the way back in 2002, and probably before. It was a common complaint of many large enterprises who felt penalised by multi-year managed service contracts that didn’t quite deliver on all the things they were promised, yet had no real means to do anything about it.
This was also an issue during the formative years of hyperscale cloud. People didn’t forget the pain they had experienced. As a result many discussions have focused on how to prevent vendor lock-in, concerned by the lack of interoperability to pick and choose solutions which were largely limited by the cloud providers’ ecosystem and service offerings.
Platformisation faces the same challenges, where financial efficiencies are weighed against functional and innovation limitations. Having worked for a hyperscale cloud company previously, the general consensus was “multi-cloud lowers capabilities to the lowest common denominator”, while customers complained “make it easier for us to do multi-cloud”. So where does the happy medium sit between these two ideas?
This is where open standards play such an important and pivotal role. Open standards are the common language that allow different software systems, hardware, and platforms to talk to one another without needing a translator. They are the antithesis of vendor lock-in and are critical for cross-platform integration for several key reasons:
Interoperability: Open standards (like IPSIE or Oauth) operate across vendors and allow customers to pick and choose which solutions they can use, without being limited to a single vendor or technology stack. Developers don’t have to reverse-engineer how a proprietary system works. If a platform supports an open standard (like Oauth for logging in), the integration path is already documented and understood.
Future-proofing and longevity: Proprietary integrations are fragile. If a vendor changes their internal code or goes out of business, the integration breaks. Open standards bring stability. Open standards are maintained by independent bodies (like the OpenID Foundation for IPSIE). They evolve slowly and deliberately, ensuring backward compatibility.
Avoiding the ‘translation tax’: Without open standards, every integration requires a custom translation layer. When two platforms speak the same open standard (e.g., two email servers using SMTP), they communicate directly. You avoid the processing overhead and potential for errors that come with converting data from one proprietary format to another constantly.
Innovation and competition: Open standards lower the barrier to entry for new competitors, which benefits the ecosystem as a whole. You can build a best-in-class tech stack. You might use a CRM from Salesforce, email from Google, and a database from Amazon. They all support open standards (like RESTful APIs), so you can stitch them together into a unified workflow.
Open standards are the fundamental bedrock of modern platformisation strategies. They shift the architectural paradigm from monolithic silos – where one vendor does everything – to modular ecosystems (where distinct, best-in-class tools connect seamlessly). This allows organisations to grow and adapt their technology stack when needed and ensures platformisation is not a one-way decision.
Politics5 days ago
Politics5 days ago
Sports5 days ago
Sports1 week ago
Fashion1 week ago
Fashion5 days ago