Robotaxi services are spreading broadly. While they’ve clearly had issues in San Francisco — as have I, driving in San Francisco — those challenges are improving. Now, both Uber and Lyft appear to be taking steps toward replacing their ride-share drivers with AI.
We may be on the verge of one of the first large-scale AI job replacement efforts that everyone will witness. This transition will fix a key issue with ride-sharing by aligning revenue more directly with its users. However, it won’t address a deeper problem that ride-sharing has in common with social media: a decoupled revenue business model — where the people using a service aren’t necessarily funding it.
Let’s talk about decoupled revenue. Then, we’ll close with my Product of the Week: a new and awesome notebook from Lenovo that sports a brand-new processor from Intel.
The Problem With Decoupled Revenue
The problem with both social media and ride-sharing is that the people paying for the service aren’t treated as the primary customers.
With social media, advertisers fund the platforms, but users engage with content as if they were the primary customers. With ride-sharing, the platform enables car owners to monetize their vehicles, but that’s a cost of the service. Since riders are the ones paying, they should be the customers of the drivers, not the platform.
In both cases, the platform prioritizes those who provide the revenue—advertisers for social media and riders for ride-sharing—while those who power the service, such as content creators and drivers, are treated as cost centers rather than valued participants.
This decoupling means that those who make the service possible — like Uber/Lyft drivers and social media users — are often overlooked. Their needs are subordinated to those of the entities providing the funding, like advertisers on social media and riders for Uber/Lyft. Instead of being seen as assets to be optimized, social media content creators and ride-share drivers are frequently treated as problems to be managed.
Both conditions are true. Users drive the advertising revenue because ads are worthless without users. Likewise, if there aren’t any drivers, there is nothing for riders to use.
Managers often fail to make this distinction. I once knew a general manager who argued that his plant would run much better without sales reps mucking things up — forgetting they were the reason the plant had business in the first place. He was fired shortly thereafter because he was an idiot, but this wasn’t a unique situation. I was an internal auditor, and it was surprising how many high-ranking executives didn’t seem to know how things actually worked — which is kind of scary when you think about it.
Businesses generally function better when the people using a service and those funding it are the same. When revenue and users are decoupled, as they are in social media and ride-sharing, platforms risk alienating the people who keep them running. You don’t mess with your revenue sources.
Will Robotaxis Disrupt or Fix Ride-Sharing?
Robotaxis may fix ride-share or badly break it. If Uber and Lyft go to robotaxi services, then the people who use this service to monetize their cars are kind of screwed unless they want to work for nearly free because they won’t be able to compete with an AI driver. However, this would also eliminate the problem the drivers represent in terms of being treated more like a cost center than a customer. With robotaxis, Uber and Lyft can just focus on the riders who then become their customers.
Now, I know these services have been considering allowing ride-share car owners to stay with the service but buy and maintain the autonomous vehicles themselves, enabling them to earn back a significant part of the revenue. That could resemble a B&B model, but with cars and trucks, maintenance, insurance, and liability concerns create additional hurdles.
Plus, while these cars will essentially be electric, we don’t have automated charging stations for them to use. Charging infrastructure is critical if autonomous vehicles operate independently, and solutions like robotic charging or battery swaps are still in development.
Given these challenges, Uber and Lyft will likely choose to own their fleets and install unique charging ports, automated charging stations, and stand-by parking lots to accommodate typical fluctuations in demand throughout the day.
Initially, There Would Be a Blend
Right now, robotaxi service technology is suitable for urban areas. However, if the rider needs to go some distance, then, at least in the short term, a human driver would still be needed until we have approved autonomous driving technology that can operate within and between cities.
We might even see some integration between autonomous cars and airplanes for service within a state. At least some of the aircraft being considered for this are VTOL (vertical take-off and landing), which provides the potential for more localized landing pads where riders could transfer from their autonomous car to an autonomous plane.
There have also been some concepts where an autonomous lifting body attaches to the car when it needs to bypass traffic or travel longer distances. That would be cool, but I think safety concerns might keep people from wanting to use it. I recall a flying car many years ago based on a Ford Pinto where the wings and car came apart at altitude, killing the company’s founders.
Wrapping Up
While robotaxis may improve ride-sharing by better aligning revenue with users, what happens to the displaced Uber and Lyft drivers? In some cases — and Tesla is exploring this — some may work remotely, taking control when a robotaxi runs into trouble to guide it to safety or a repair station. But this would require far fewer drivers than today, offering only temporary employment for some, while most would be out of luck.
As AIs move in to take over more of these jobs, figuring out what to do with the people who have lost theirs as a result will be an increasingly complex problem. What happens to drivers, or even pilots, when vehicles can drive or fly themselves? It isn’t like you can suddenly pivot all of them to robot trainers. We don’t need nearly as many trainers as we now have drivers and pilots.
For us users, things will become much more convenient with lower prices, better service, and less concern about whether your driver is a serial killer. (I watch too many police TV shows!) There will be fewer interesting conversations, although, come to think of it, generative AI is conversational, so you might be able to talk to the car during your ride, and it will certainly be less likely to be abusing drugs or reading messages instead of focusing on the road.
The experience should be cheaper, safer, and far more convenient, which will fix the decoupled revenue problem.
Lenovo IdeaPad Pro 5i Prototype
The new Lenovo IdeaPad Pro 5i prototype is pretty awesome. The shipping version of this laptop uses Nvidia RTX graphics, but this one uses Intel ARC and isn’t due until July.
This AI PC, one of Intel’s first, uses the 9u processor — also called the Ultra 9 processor. It has a 16-inch OLED screen that looks awesome, and OLEDs have become more efficient, so they aren’t the huge battery killers they once were.
I prefer a larger screen because I generally work off a desktop machine, so going down to a 14-inch or even 13-inch is incredibly painful. Fifteen inches is fine, but that extra inch on a 16-inch makes a noticeable and positive difference.
The downside is I can’t really work on this laptop on a plane. Still, I generally just watch movies when I’m flying because I need time to unwind and chill out since I just don’t feel like working every single moment anymore, and films on this screen are incredible.
When this laptop ships, the price should be around $1,500 (this could change with tariffs), which is reasonable for a 16-inch OLED with discrete graphics. Intel ARC graphics use a blend of AMD’s and Nvidia’s approaches to upscaling, and, in their stand-alone cards, ARC represents one of the better values in GPUs. The IdeaPad Pro 5i is the first laptop I’ve seen with these graphics.
The drivers on this prototype were undergoing some changes, but I’d expect between 10 and 20 hours of battery life depending on use, which would be huge for a 16-inch laptop.
The graphics are more than adequate for my needs. There still isn’t a lot of desktop AI stuff outside Office 365. So, while its NPU is light, with 11 TOPS compared to over 40 for other AI PCs, the total TOPS is close to 100, with the GPU contributing additional AI processing power.
The downside is that this laptop won’t run Recall or Cocreator until Microsoft enables those. Recall is on hold and only available in trial form, although I’ve found ChatGPT’s Dalle-E web implementation to be just fine. (Microsoft could enable Dall-E on laptops with discrete graphics; it’s only a matter of time). Given how fast AI is advancing, it is nice to have an AI PC from Intel because most companies still test Intel hardware, which is better to ensure a positive outcome.
Overall, this is a very nice product. Mine, pictured above, came in silver-black. It is attractive, useful, and well-priced, making the Lenovo IdeaPad Pro 5i prototype with ARC graphics my Product of the Week. Sorry — you won’t be able to buy one for yourself until July.
