• TfL introduces radar cameras that monitor five lanes without visible alerts
• Half of London’s 2024 fatal collisions involved excessive speed
• Cameras will be installed on 20mph and 30mph roads across ten boroughs

Transport for London (TfL) is moving ahead with trials of radar-based speed cameras which differ significantly from existing roadside systems in both design and operation.

The new devices combine 4D radar tracking with 4K imaging, removing the need for embedded road sensors, visible flashes, or painted markings that typically signal enforcement zones to drivers.

A legal feud between the co-founders of Lux Optics, the developer behind the Halide camera app, revealed that Apple was close to acquiring the company. As first reported by The Information, Apple held acquisition talks for Lux Optics, which also developed the Kino, Spectre and Orion apps, in the summer of 2025.

According to The Information, the deal eventually fell through in September of that year, but the potential acquisition could’ve provided Apple with the third-party software to improve its own built-in camera app. Apple is already rumored to be introducing variable aperture to its upcoming iPhone 18 Pro models, so it’s not surprising that the iPhone maker was looking for software with advanced features to match its possibly upgraded camera hardware.

Despite Apple’s interest, Lux Optics’ co-founders, Ben Sandofsky and Sebastiaan de With concluded that future updates to Halide could increase the company’s valuation and ended the acquisition talks. According to the lawsuit between the co-founders, Sandofsky started investigating de With for the alleged misuse of company funds shortly after the talks with Apple ended. Afterwards, de With was fired from Lux Optics and later joined Apple’s design team. While Halide may remain third-party software for iPhones and iPads, users can still look forward to some software improvements to the built-in camera app, since that’s reportedly one of Apple’s priorities.

This week, a topic that has been boomeranging around Silicon Valley bounced into the spotlight: AI tokens as compensation. The idea is straightforward enough — rather than giving engineers only salary, equity, and bonuses, companies would also hand them a budget of AI tokens, the computational units that power tools like Claude, ChatGPT, and Gemini. Spend them to run agents, automate tasks, crank through code. The pitch is that access to more compute makes engineers more productive, and that more productive engineers are worth more. It’s an investment in the person holding them, is the idea.

Jensen Huang, the leather-jacket-wearing CEO of Nvidia, seemed to capture everyone’s imagination when he floated the notion at the company’s annual GTC event earlier this week that engineers should receive roughly half their base salary again — in tokens. His top people, by his math, might burn through $250,000 a year in AI compute. He called it a recruiting tool and predicted it would become standard across Silicon Valley.

It isn’t entirely clear where the idea was first, well, ideated. Tomasz Tunguz, a renowned VC in the Bay Area who runs Theory Ventures and focuses on AI, data, and SaaS startups — and whose writing on all things data has garnered a loyal following over the years — was talking about this in mid-February, writing that tech startups were already adding inference costs as a “fourth component to engineering compensation.” Using data from the compensation tracking site Levels.fyi, he put a top-quartile software engineer salary at $375,000. Add $100,000 in tokens and you’re at $475,000 fully loaded — meaning roughly one dollar in five is now compute.

That’s no coincidence. Agentic AI has been taking off, and the release of OpenClaw in late January accelerated the conversation considerably. OpenClaw is an open-source AI assistant designed to run continuously — churning through tasks, spawning sub-agents, and working through a to-do list while its user sleeps. It’s part of a broader shift toward “agentic” AI, meaning systems that don’t just respond to prompts but take sequences of actions autonomously over time.

The practical consequence is that token consumption has exploded. Where someone writing an essay might use 10,000 tokens in an afternoon, an engineer running a swarm of agents can blow through millions in a day — automatically, in the background, without typing a word.

By this weekend, the New York Times had put together a smart look at the so-called tokenmaxxing trend, finding that engineers at companies including Meta and OpenAI are competing on internal leaderboards that track token consumption. Generous token budgets are quietly becoming a standard job perk, the paper reported, the way dental insurance or free lunch once was. One Ericsson engineer in Stockholm told the Times he probably spends more on Claude than he earns in salary, though his employer picks up the tab.

Maybe tokens really will become the fourth pillar of engineering compensation. But engineers might want to hold the line before embracing this as a straightforward win. More tokens may mean more power in the short term, but given how fast things are evolving, it doesn’t necessarily mean more job security. For one thing, a large token allotment comes with large expectations. If a company is effectively funding a second engineer’s worth of compute on your behalf, the implicit pressure is to produce at twice the rate (or more).

And there’s a muddier problem underneath that: at the point where a company’s token spend per employee approaches or exceeds that employee’s salary, the financial logic of headcount starts to look different to its finance team. If the compute is doing the work, the question of how many humans need to be coordinating it becomes harder to avoid.

Jamaal Glenn, an East Coast-based Stanford MBA and former VC turned financial services CFO, similarly points out that what may seem like a perk can be a clever way for companies to inflate the apparent value of a compensation package without increasing cash or equity — the things that actually compound for an employee over time. Your token budget doesn’t vest. It doesn’t appreciate. It doesn’t show up in your next offer negotiation the way a base salary or equity grant does. If companies successfully normalize tokens as pay, they may find it easier to keep cash comp flat while pointing to a growing compute allowance as evidence of investment in their people.

That’s a good deal for the company. Whether it’s a good deal for the engineer depends on questions most engineers don’t yet have enough information to answer.

The app also includes access to two scheduled operational modes for those who would like to leave the robot in the pool, including a calendar-based mode with three frequency levels—90 minutes x 2, 60 minutes x 3, or 45 minutes x 4. The other mode is a bit of a letdown: The so-called AI Navium mode sounds like it uses the AI camera to periodically survey the pool over the course of a week and perform a routine cleaning only when required—but in reality, this mode merely performs a quick analysis of your previous runs and then uses AI to create a schedule for the next few days, based on how you’ve used the robot in the past.

Hungry for Gunk

The Scuba V3 made fairly quick work of debris in my pool during test runs, rarely needing more than a couple of hours to scoop up all visible detritus on the pool floor while also scrubbing the walls and waterline. The AI camera system does seem to work as advertised, even locating small pebbles I tossed into the pool and dutifully routing itself to collect them. With organic debris, the pool looked fully clean after each run (ending between 170 and 190 minutes each time), and with synthetic debris, the Scuba V3 achieved a 96 percent cleanliness rating, with just a few test leaves remaining in some difficult corners. That’s especially good performance given that three hours is not a lot of operating time. And note there’s no way to adjust the running time outside of the scheduled modes; on-demand modes always run the battery until it’s nearly dead. Fortunately, Aiper does seem to make the most of this time, formally specifying a maximum coverage area of a significant 1,600 square feet.

I unfortunately didn’t have much success with the AI schedule mode. After running the analyzer, the app suggested a baffling five-day schedule comprising two floor runs, two floor-plus-waterline runs, and a final floor run. It then ignored the schedule and promptly ran a three-hour floor run, which drained the battery completely. I tried again the next day, and the robot missed its schedule, then ran randomly late in the night. I wasn’t a big fan of leave-it-in-the-pool scheduling before testing the Scuba V3, and this showing didn’t improve that opinion.

When finished with a run, the Scuba climbs to the waterline and sends a push notification to the app, alerting you that it’s ready to be collected and cleaned. Note that you only have 10 minutes to reach it: The Scuba can’t float, so it has to use the last of its juice to run a motor to tread water and hold itself in place. After that 10 minutes is up, the spent Scuba sinks to the floor of the pool and must be retrieved with a pool and hook. My best advice is to set a 175-minute timer each time you launch a run to remind you to watch for the completion notification.

Cleanup can be somewhat involved. The filter basket design features a large lid that makes it easy to access the inner filter, and hosing down both of these filters clean is straightforward. The removable mesh on the interior basket is another story, though. While it’s very effective at capturing dirt and other very fine debris, it’s quite difficult to clean, and if you don’t remove it from the basket, lots of debris gets caught between the mesh and the basket itself. Removing and replacing the mesh is difficult, especially when it’s wet, so I usually just left it in place and cleaned it the best I could after each run, accepting that it would never be perfect. I expect most users will do the same.

If you thought Apple accessories were getting expensive, Hermès has just taken things to a completely different level.

The luxury fashion house is now selling a range of MagSafe-compatible chargers priced from $1250, with some models going well beyond that price.

At the entry point, the Paddock Solo Charger is a single-device magnetic charger priced at $1250. If you step up to the Paddock Duo at $1750, you can charge both an iPhone and an Apple Watch at the same time. Furthermore, there’s also the Paddock Yoyo, also $1750, which adds a wraparound USB-C cable designed for travel.