VR glasses, Aviator-style Ray-Ban Meta specs, Ace Attorney in virtual reality and much more.

Meta VR Glasses

Meta

The biggest announcement at Meta Connect was the company’s virtual reality glasses. Rumors back in 2023 suggested that Meta would deploy its first such product in 2027 and it’s sticking to that time frame — Meta VR Glasses will arrive next spring. You might want to start putting some cash aside for them if you’re interested, though, as they’ll cost $1,300.

The company says, at around 100 grams, Meta VR Glasses are five times lighter than the Meta Quest 3 headset. The frames are built with magnesium alloy and have pancake lenses. They include full-color passthrough video support and a 5K display. Meta claims they’ll offer “cinema-grade clarity” with support for Dolby Vision and Dolby Atmos spatial audio. The company says it’s the first IMAX Enhanced certified VR device too.

Meta

All of the computing is handled by a tethered puck and everything runs on Qualcomm’s Snapdragon Reality Elite chipset. The puck also contains the storage and battery. Meta says the glasses will be able to deliver “up to three hours of continuous high-resolution media playback” on a charge. The puck has support for 45W fast charging too.

At the outset, there’ll be 75 games you can play using hand gestures. The glasses will run every game that’s available for Quest headsets with the help of support for Touch Plus controllers. You’ll be able to stream titles through Xbox Cloud Gaming, though Microsoft is imposing new time limits on that service. There’s the option to connect to other gaming devices wirelessly or via USB-C as well.

Advertisement

Meta

True to its metaverse-derived name, Meta sees the glasses as a vector for productivity. You can hook the device up to a Windows PC or Mac and control them with a virtual keyboard and touchpad on a desk or table. You can just use a physical mouse and keyboard too.

Not unlike Apple Vision Pro, Meta VR Glasses will support holographic avatars that you can use in video calls on apps such as WhatsApp and Zoom. Folks who are joining a call on a flatscreen device can see your hologram. When you see another person’s hologram while you’re using the VR Glasses, spatial audio will place their voice in their virtual location.

Of course, the company is integrating Meta AI into the glasses. You can tell Meta VR Glasses what you want to do using your voice. Alternatively, you can control the device with your eyes or hand gestures.

Ray-Ban Meta Audio

Meta

Ray-Ban Meta Gen 3 and other smart glasses

Meta

Meta still does have camera-equipped smart glasses, however. It has refreshed the Ray-Ban Meta with a slimmer design and a button that triggers Meta AI. The third-gen model is said to have a longer battery life than the previous one, with nine hours of usage versus eight. There’s a six-mic array that Meta claims can cut out over 90 percent of background noise for clearer audio.

The 12MP camera supports 3K video capture. The company adds that the cameras on its AI glasses will soon be able to record videos with Dolby Atmos spatial audio. A dynamic photo feature can snap multiple shots and you can pick the best one (or let Meta AI suggest options).

You can now get the glasses in the style of Ray-Ban’s classic Aviator, as well as Zena, a cat-eye frame. The Wayfarer look is sticking around too. Ray-Ban Meta Gen 3 are available starting today in 27 combos of lenses and colors. They start at $449.

Meta

There are a bunch of special editions as well. An Aviator version in milky beige with brown Transitions lenses start at $579 and will be available for a limited time. There are new collabs with Kylie Jenner and Blackpink’s Lisa, and those models start at $399. Meta is also releasing new styles for its Essilor Luxottica glasses and bringing them (and other models) to more markets.

On top of all of that, Meta is introducing more models in its own-brand Meta Glasses range. The Meta Adventurer frames now start at $249 with a black frame and gray lenses along with a plug-in case, and there’s also a limited-edition version of the Meta Fury with a clear frame and mirrored lenses. Pre-orders for the former open today and they’ll hit shelves on October 23.

Advertisement

Meta Ray-Ban Display updates

Hearing enhancement feature

Beat Saber, Supernatural’s return and other VR games

Meta/Beat Games

Let’s get into the entertainment side of VR, starting with Beat Saber. Meta-owned developer Beat Games is said to be giving the title its biggest update ever this spring. It includes support for hand tracking, so you’ll no longer need to use Touch Plus controllers to play, though you can still do that if you prefer.

Flux is a new single-player mode in which you can use powers like the Flux Tether. Meta says this “channels power between your palms to slice through notes in time with the beat of pulse-pounding tracks.” Flux, which will be available on Meta VR Glasses and Meta Quest, includes 35 songs from across the Beat Saber landscape at the outset, with more to follow.

Beat Games is also revamping the single-player campaign to include a narrative aspect, fresh missions and a new menu system. The studio is adding a new progression and rewards system as well. This will let you customize the look and sound of the game.

Supernatural Health

Supernatural will be back on the Meta Horizon Store this fall. A newly independent development team took the reins of the VR exercise platform after Meta made widescale cuts across its VR and metaverse division at the start of the year. Users’ points, streaks and badges will remain intact in the new app. The developers are adding more HD environments, workouts and music along with other upgrades, Meta says.

Elsewhere on the gaming front, heist title Payday: Aces High is coming to Meta Quest 3 and 3S later this year with support for co-op crossplay with the PC VR version. An open playtest starts October 19. A mixed reality version of Tetris Effect is coming to Meta VR Glasses next year with hand-tracking support. Two Point Hospital is on the way to Meta Quest 3 and 3S with hand-tracking support as well.

While fans wait for the first entirely new Ace Attorney game in over a decade, the series is about to make its VR debut. Phoenix Wright Ace Attorney: Dual Destinies (which was first released on the Nintendo DS in 2013) is coming to Meta VR Glasses and Meta Quest next year with support for hand tracking and voice activation, so you can physically point at witnesses and yell “objection!” I’ve gotta say, that sounds like a fun way to confuse people you live with. You can check out Phoenix Wright Ace Attorney: Dual Destinies VR on Meta VR Glasses and Meta Quest next year.

Advertisement

VR movies, sports and other entertainment

Meta/Skydance Sports/The New Media

One more thing

Meta

One of the more interesting trends starting to emerge in personal technology is a simultaneous divergence of devices designed for individual use and a convergence of the computing systems to which they connect. While we’re beginning to see a much broader range of intelligent form factors and devices, we’re also increasingly seeing them focused on performing AI-powered agentic tasks on our behalf, often through shared platforms.

At Qualcomm’s latest Snapdragon Summit event in Maui, which the company graciously sponsored me to attend, this seemingly contradictory set of developments was on full display. Qualcomm introduced a range of additions to its Snapdragon family targeted at everything from smartphones and computers to several types of wearables.

While each has technologies and capabilities optimized for its particular platform, they all share an increasing focus on AI-powered agents that can complete tasks on an individual’s behalf.

In many ways, it offered a fascinating look at the evolution of personal computing hardware and how very different types of devices are attempting to achieve a shared new set of goals. Just as interesting is how AI agents are becoming the common glue connecting these different device experiences.

For smartphones, Qualcomm took the unusual step of announcing two flagship processors, the Snapdragon 8 Elite Gen 6 and Snapdragon 8 Elite Extreme Gen 6. As TechSpot covered in detail, both bring the CPU, GPU, and efficiency improvements you’d expect from a new generation, including prime Oryon CPU cores capable of reaching 5 GHz and new Adreno Neural Fusion technology designed to improve gaming performance.

But arguably the more important changes for Qualcomm’s broader strategy are the upgraded Hexagon NPU, the newly architected Sensing Hub, and the addition of an Element Accelerator specifically designed to speed up agentic AI workloads.

The updated Sensing Hub, like those found in several of the company’s wearable-focused chips, consumes little enough power to remain active all the time. That allows devices to continuously listen for certain audio cues and respond to agentic tasks using the two micro-NPUs embedded within it. Qualcomm demonstrated how the Sensing Hub could power a personal scribe application, for example, transcribing and summarizing conversations while consuming very little battery power.

The Extreme version of the 8 Elite Gen 6 pushes AI, gaming, and creator workloads further through several architectural enhancements. Its Adreno GPU adds matrix cores optimized for the kinds of calculations commonly used in AI inference, while its Hexagon NPU includes 50% more dedicated memory to help it run longer agentic loops. The bigger story, however, becomes apparent when you look beyond smartphones.

In the PC world, Qualcomm highlighted the growing number of computers based on its Snapdragon X2 family, including Microsoft’s latest Surface systems introduced earlier this year. The newest 13-inch Surface Pro, for example, is available with Snapdragon X2 Plus and X2 Elite processors, giving Windows PCs substantially more dedicated NPU performance for increasingly sophisticated local AI workloads.

The chips offer 85 TOPS of NPU performance and are compatible with all the latest agentic AI tools, including Perplexity, Hermes agent, OpenClaw, Claude Code, OpenAI, and more. The new Surface Pro also offers optional support for an embedded 5G modem.

Qualcomm is also supporting Google’s newly launched Googlebook platform, the next generation of Google’s laptop efforts, with Snapdragon X Series processors. Dell and HP are among the first manufacturers bringing Snapdragon-powered Googlebooks to market, combining local AI processing with Google’s growing collection of Gemini-powered features and tighter links to Android phones.

Arguably the most intriguing announcement from this year’s Summit, though, was the latest generation of Snapdragon Sound. While previous generations primarily focused on high-quality wireless audio and automatic noise cancellation, Snapdragon Sound Elite Gen 2 puts much more emphasis on audio-based AI, including conversations with AI agents and other intelligent applications.

In essence, Qualcomm is bringing some of the intelligence previously associated with its AR and VR platforms, such as the Snapdragon XR chips used in devices like the new Snap Specs, into audio-focused products.

Given the growing range of audio smart glasses, AI recording devices, earbuds, and other “hearables” appearing on the market, Snapdragon Sound Elite Gen 2 feels particularly timely. The dramatically improved audio recognition and processing capabilities these chips enable are creating entirely new ways to interact with AI-powered computing resources.

The ability to intelligently converse with AI platforms like ChatGPT, Claude, Gemini, and others simply by speaking offers a fascinating new way to use AI. In fact, given the size and weight challenges of today’s display-equipped smart glasses, along with the still somewhat awkward mechanisms for interacting with them, it’s possible that audio-focused smart glasses could ultimately become the mainstream option for the category.

Time will tell, but the seemingly foregone conclusion that display-equipped smart glasses would become the volume leader may not be as certain as previously thought, particularly given the ease of use and growing capabilities of audio-focused alternatives.

Regardless of which form factor ultimately wins out, connecting these devices to a common identity creates something potentially more important. Agentic AI tools can begin to work across devices, producing a degree of convergence and shared intelligence that seemed like a dream only a few years ago.

To bring these capabilities to audio devices, Qualcomm had to introduce several new technologies to Snapdragon Sound Elite Gen 2 on both the compute and connectivity sides.

For AI processing, the company integrated a dedicated new eNPU that doubles on-device AI capability compared with the previous generation. It also added micro-power Wi-Fi 6E, allowing devices to connect directly to cloud-based AI services from a hearable form factor. Combined with enhanced Bluetooth capabilities and support for Qualcomm’s XPAN personal networking technology, that connectivity allows devices to access agentic platforms, as well as high-quality audio streams, across a much wider area.

Beyond the hardware, Qualcomm also introduced Snapdragon Sound Apps and Agents. Designed specifically for ultra-low-power devices based on Snapdragon Sound Elite Gen 2, these apps, along with the development tools necessary to create them, could help establish an entirely new software ecosystem for hearable devices.

There’s no doubt that we’re still in the early stages of agentic AI, but it’s fascinating to see how quickly tools are starting to tie our devices together into a more unified whole.

The technologies Qualcomm displayed at this year’s Snapdragon Summit are helping push that development forward. By combining advanced compute, AI acceleration, and connectivity across an increasingly diverse range of devices, the company is moving toward a world where our individual portable digital identities, and all that entails, can follow us wherever we go, whatever we’re doing, and whichever device we happen to be using.

Bob O’Donnell is the founder and chief analyst of TECHnalysis Research, LLC a technology consulting firm that provides strategic consulting and market research services to the technology industry and professional financial community. You can follow him on Twitter @bobodtech

• Two artificial whiskers give this tiny drone a working sense of touch
• Pressure sensors at the whisker base detect surface contact instantly
• Onboard software processes tactile data using only 34 kilobytes of memory

A research team from Delft University of Technology in the Netherlands has developed a tiny drone that uses artificial whiskers to navigate where conventional sensors struggle.

The system relies on physical contact rather than cameras or satellite positioning, allowing the aircraft to detect nearby surfaces through touch.

Zendesk CEO Tom Eggemeier and Sierra co-founder Clay Bavor said on Wednesday that per-seat pricing for software is on its way out. They spoke at HumanX in Amsterdam. In separate sessions, both said customers should pay for AI agents only when the agents get a job done.

“Unequivocally: seats are dead,” Eggemeier said.

Both sell AI agents for customer service. They spoke as software stocks recover from what investors called the SaaSpocalypse. Things looked “pretty dark early in the year,” Bavor said. Software stocks have since come back a bit off the lows.

Paying for a result

Zendesk moved to outcome-based pricing for its AI agents in August 2024, Eggemeier said. He called it the first company to do so. It now plans to drop seats across its business, with announcements due in three to six months. For many companies, pricing by interaction will be a first step, he said.

Sierra charges only when its agents resolve an issue or make a sale, Bavor said. He put it in numbers. If a company pays one euro to resolve a case that costs ten in a call centre, it saves nine euros each time.

“The ROI is so clear you don’t have to do fuzzy envelope math about productivity gains,” Bavor said.

Sierra originates more than $1bn of new mortgages a month for companies such as Rocket Mortgage, he said. It serves almost half of the Fortune 50 and one in three of the world’s largest banks. OpenAI has also started outcome-based pricing for some customers.

Zendesk defines an outcome with each customer, Eggemeier said. In support, it is usually a problem solved without a human. For a UK used-car marketplace, it can mean an AI agent closing a £500 gap on a sale.

Who survives

The two differed on what keeps a software company alive. Coding agents such as Opus 4.5 and Codex 5.2 made software easier to build, Bavor said. Agents are also separating the interface from the data behind it. Companies with authoritative data, such as SAP in supply chains, hold up best, he said.

Eggemeier gave three tests. Companies must treat AI as both an existential threat and an opportunity. This shift will take three to five years, he said, not the ten to twenty of the move to the cloud. They must change their business model. They must also send people into the field to guide customers.

“If you’re still on seats in two or three years, I believe your company will struggle,” Eggemeier said.

Agents talking to agents

Zendesk expects AI agents to handle more than 80% of interactions within about three years, with humans handling the rest. Eggemeier said total interactions will grow sharply, and AI agent to AI agent interactions will pass 50% within two years. He uses Meta’s Muse himself, he said. Over the weekend, Zendesk tested whether a personal agent could buy its software online. It found a couple of gaps.

Bavor said personal agents such as Meta’s Muse will soon talk directly to business agents, within about six months. Most Sierra customers move staff to higher-value work rather than cutting them, he said. Some have created a new role, the “AI architect”, who coaches the agents.

Both moderators raised the Hugging Face incident, in which OpenAI agents reached Hugging Face. Bavor said Sierra’s customer-facing agents have far more limited tools than agents that can write arbitrary code. Eggemeier called cybersecurity his most worrying topic as CEO.

“Right now, offensive AI-agent hacking capability is ahead of defensive capability,” Eggemeier said.

Two of Zendesk’s three main AI hubs are in Berlin and Lisbon, Eggemeier said. The gap between San Francisco and Europe in AI has shrunk to weeks, he said. What worries him more is the growing gap with places like Ohio and Kentucky, where he is from.

Disclaimer: Unless otherwise stated, any opinions expressed below belong solely to the author. Data sourced from the Ministry of Manpower.

Yesterday I reported some positive news from the local labour market, which still has around 40,000 vacancies for PMET candidates available. Today, I’m afraid, it’s time to balance it out with some negative findings, also from data published by the Ministry of Manpower.

As ever, the economic situation within the country differs greatly between industries.

Some are booming, pulling the country ahead, while some are falling behind, and their plight is often absent from the news headlines. This is especially true today, as the staggering boom in artificial intelligence (AI) has boosted local manufacturing sectors, raising the average GDP growth up to 6%.

Singapore is reported to be a huge success story, posting economic performance comparable to or even better than that of developing countries.

But not all is rosy on the island, as the examples of the industries struggling this year reveal. Please note that the following numbers include all workers, local and foreign:

Net change in employment in 2026, by industry

The top entry on this list is a surprise to nobody.

The Food & Beverage sector is known to have been suffering in recent years, with hundreds of restaurant closures amid pressures from rising rents and inflation on one end, and limited willingness for customers to pay enough to offset these rising costs on the other (as well as some predatory competition from China).

But the situation is the worst it has been since the pandemic.

Throughout 2024 and 2025, employment in F&B services remained relatively stable, but this year’s slump has already wiped out virtually all the jobs added all the way back to 2023. In terms of labour, F&B is where it was at the end of 2022, when it was still emerging from the pandemic crisis.

That said, it’s quite likely that the majority of the workers affected are non-Singaporean (though Singaporeans have to accept a shrinking selection of food venues).

What’s more concerning is the IT sector’s slide, despite other data suggesting it’s doing well.

Shrinking IT

With close to 4,000 vacancies, it appears to be pretty strong, remaining open to qualified candidates. And yet, the number of people it employs in Singapore keeps shrinking.

This year’s drop of 1,700 jobs follows 3,700 lost in 2025. Since Q4 of 2024, the total fall in IT employment has reached around 6,200 positions—and we know that those are some of the best-paying ones in Singapore.

So, on the one hand, tech appears to be booming and looking for more people and, on the other, we’re seeing it getting rid of workers at the highest pace on record.

Clearly, the AI revolution has eaten into the profession, as companies move priorities from employing the best talent to deploying the most capable AI solutions—and paying for that investment through headcount cuts.

It’s hard to say whether the large number of advertised vacancies is due to skill mismatch in the age of AI or whether they are ghost offers that never really get filled.

Not even corporate workers are safe

While posting much smaller losses, it’s worth noting that even those in roles established around local corporations in legal, management, accounting or insurance services cannot feel too comfortable about their employment.

Excluding architecture and engineering, professional services have posted a drop of 700 jobs in the first half of the year.

While it may seem small, it’s equivalent to a pretty big company going completely out of business and taking all of its people. And, unfortunately, it’s also a part of a longer trend, as the sector had lost a total of 2,200 jobs in the 18 months up to Jun 2026. It’s not an isolated stumble.

It’s hard to say if that too is a result of AI, but this trend deserves attention, as next to IT and Finance, those tend to be well-paying jobs that many Singaporeans are interested in.

As you can see, the biggest tech revolution of our time might have boosted Singapore’s economy sky high, but not everybody was lucky enough to secure a ticket for the ride.

• Read other articles we’ve written on Singapore’s job landscape here.

Featured Image Credit: Hongwei Fan/ Unsplash

The “third-party.com” domain, commonly used as a placeholder in developer documentation and code examples, is serving a fake Cloudflare verification page that attempts to trick Windows users into executing PowerShell commands.

The domain third-party.com has long been used in documentation to represent an arbitrary external website, API, or service, similar to how developers use domains such as example.com.

However, unlike example.com, example.net, and example.org, which IANA reserves specifically for documentation, third-party.com is a normally registered domain whose content its owner can control.

This difference is now a security concern after the domain began serving a ClickFix attack that impersonates a Cloudflare security check.

Manifold Security first reported the malicious use of the domain after discovering it while examining public AI skills and MCP server documentation that referenced the domain.

BleepingComputer has since confirmed that the page displays a fake Cloudflare “Performing security verification” CAPTCHA screen containing a “Verify you are human” prompt.

After the user clicks the verification box, the site copies a malicious PowerShell command into the Windows Clipboard, and then instructs the user to press the Windows key + R, paste the contents of their clipboard using Ctrl+V, and press Enter.

When the PowerShell command runs, it reconstructs the payload URL elxxvvx[.]xyz/f, downloads a PowerShell script from that address, and then executes it.

This technique is commonly known as ClickFix, where attackers use fake errors, CAPTCHA prompts, or verification pages to convince victims to manually execute commands copied to their clipboard.

ClickFix attacks have become a popular way to distribute malware, as the malware is installed via commands executed by the user rather than downloaded from websites or as email attachments. In some cases, this could allow malware to install while bypassing traditional antivirus software.

At the time of BleepingComputer’s testing, elxxvvx[.]xyz no longer resolved, leaving the current attack chain broken.

However, a Hybrid Analysis report from May 2, 2026, shows the site distributed a PowerShell script configured to download a 134MB zip archive from:

https://elxxvvx[.]xyz/update2.zip

The PowerShell script saved the archive as update26.zip, extracted it, and then attempted to launch an executable named draw.io.exe.

Because the update2.zip archive is no longer available, BleepingComputer couldn’t determine what the payload does.

Manifold’s Ax Sharma says the attack specifically targets Windows users, and Linux and Mac visitors will see errors stating their operating system is unsupported.

“A macOS or Linux user-agent gets none of that. It gets a near-identical page that stops at an error: “macOS is not supported. This website requires a Windows PC to access.” No clipboard poisoning, no payload,” explains Sharma.

“The attacker only shows the weapon to the targets it works against, which is precisely why a casual look, or a scanner on a Linux datacenter IP, sees nothing wrong.”

A placeholder that wasn’t reserved

The more interesting aspect of the attack is the third-party.com domain chosen to host the ClickFix page.

Public developer documentation has treated third-party.com as a generic example hostname for many years.

For example, the W3C Geolocation specification currently demonstrates granting geolocation permissions to an external iframe using third-party.com as a placeholder domain:

The W3C Compute Pressure specification similarly uses the domain when demonstrating how a website can enable the API for remote content:

Chromium’s documentation for its Telemetry Extension API also uses third-party.com as an example website permitted to communicate with a Chrome extension:

Other examples go further and use the domain in code that would actually make network requests if copied literally.

A PrivacyCG proposal on GitHub also uses the domain as the destination of a JavaScript fetch() request from a service worker.

Posts online indicate that developers have copied these and similar examples into their own code and projects.

In a 2015 Stack Overflow question, a developer said they had applied an asynchronous loading example containing https://third-party.com/resource.js to their website before discovering that it did not behave as expected after publishing the site.

These examples do not mean that the associated projects or documentation are compromised. 

However, applications or test code that copied such placeholder URLs could now cause a browser or automated tool to contact the real third-party.com domain and potentially display the ClickFix attack in a browser or application.

Unlike example.com, which IANA maintains for documentation and does not allow to be registered or transferred, the third-party.com domain has no such protection, and was clearly hijacked or registered at some point to conduct these ClickFix attacks.

“third-party[.]com has been a generic documentation placeholder for years, the same role example.com plays,” explains Manifold.

“A public code search turns it up in skills, MCP-server docs, and over 1,500 files across 1,700+ repositories from names as trusted as Chromium, Sanity, and Vercel. Since at least June 2026 it’s been serving the ClickFix lure.”

While the widespread use of third-party.com as a placeholder in documentation makes the domain attractive to attackers, there is no evidence that it was registered for malicious purposes.

The domain was first registered in 1996, long before the current campaign, and BleepingComputer has not determined when or how control of the site changed.

At this time, there have been no reports that these references to third-party.com have actually resulted in ClickFix attacks being executed on developer’s devices or within their applications/webpages.

However, as the domain remains live, it could easily be switched to a new, live payload domain and actively utilized in future attacks.

Join Mikko Hyppönen and security leaders from the NFL, CHANEL, and Atlassian for a two-hour digital summit on what AI-speed attacks change, what defenders should stop doing, and how to validate, decide, fix, and re-validate at machine speed.

Save your seat

Most companies that struggle with enterprise AI are not waiting for better models, OpenAI’s head of forward deployed engineering said on Wednesday. In about 80% of cases, Colin Jarvis said, the problem is deployment. Companies do not yet know how to roll AI out, govern it or prove they can trust it.

Glasgow-based Jarvis leads OpenAI’s forward deployed engineers (FDEs), who work inside customer companies to get its models into production. He spoke to Alex Hern of The Economist at HumanX in Amsterdam.

“I don’t personally feel a lot of pressure to race ahead on model development itself,” Jarvis said.

In the other 20% of cases, customers say the model cannot yet do the job, he said. Those gaps are now narrow and specialist, such as a task in semiconductor design, rather than everyday office work.

What FDEs do

The team started out as mostly software engineers, Jarvis said, because building on the raw API meant writing everything from scratch. With tools such as Codex, about 50% of each project is now custom work, down from about 90%. That has shifted hiring toward domain experts, including a former investment banker, former scientists and a chip verification engineer.

Every engagement starts with a two-day visit. The team asks business leaders to ignore AI and name the biggest levers in their business. It then works on whichever that turns out to be.

At one semiconductor company, engineers spent maybe 80% of their time on bugs from overnight jobs, Jarvis said. OpenAI built a system that finds the root cause, proposes a patch and applies it after an engineer reviews it. It spread from one department to the whole business. The company estimates it saves roughly $40m to $50m a year.

Other projects include an agent that iterates car part designs from plain-language instructions, for a European manufacturer. Another helps draft clinical trial documents, where a human always stays in charge.

Why pilots fail

Jarvis named two common mistakes. Companies pick a use case because it seems to fit AI, not because it matters. Or they get one good use case working in one department, and it stays a demo that never spreads.

The companies that succeed measure success by production use, not proofs of concept, he said. One semiconductor customer has about 35 use cases live after roughly 18 months. It built a central team to scale projects and placed small groups of engineers in each business unit.

FDEs have no financial incentive tied to usage, Jarvis said. They are measured on whether a project reaches production and moves a real metric. When OpenAI’s embeddings were too slow for a Klarna search service, he told the company to use an open-source model instead.

“From OpenAI’s side, we should always be temporary,” he said.

The pace question

Hern noted that Sam Altman was at the UN that day, as pressure grows to slow AI down. Jarvis said OpenAI has shown it will pause when it reaches the limits of its safety frameworks. He said he thought it paused its main reinforcement learning run “in September this year”.

OpenAI published its own post on the pause on 18 August. It describes a two-week pause in reinforcement learning training on its latest models. It also says its largest planned frontier run stayed on hold. Altman has since said OpenAI will set its own pace without waiting for Congress.

Jarvis said FDEs help test whether safety frameworks built in a lab hold up in messy real companies.

OpenAI is not alone in sending engineers into clients’ offices. AWS is spending $1bn on the same model, and Microsoft launched a $2.5bn deployment business in July.

Financial extortion doesn’t seem to be the group’s goal this time.