Shimano, Japanese maker of bicycle gears and other components has rather got left behind in the surge of popularity of eBikes.
After COVID, there has been a drop in bike sales and they have been using their excess resources in developing an AI gear set, the Q’Auto for a conventional non-eBike.
This will learn from a cyclist’s style of riding, the terrain they cycle on and in conjunction with pre-loaded data, will select a suitable gear automatically for the cyclist.
One test user, a women in here forties, commented:
“I can’t believe how easy it was to go up a hill,” she said. “I don’t think I would need an [electric bike].”
Sounds pretty crazy to me, I guess the quote must mean that there are some riders who’ve just never worked out how to use gears. More importantly, they must think they can tap into the “I spend more time buying high-cost accessories and extras for my bike than riding it” market.
We get some guests here who come with 10k+ mountain bikes, but they’re using them on the World-Cup level downhill courses, so it’s the right tool for the job. But those people apart, how many of these massively expensive bikes, and road race ones for that matter, are ever used for something that all the extra cost will actually benefit?
As for gears, I’m the type of person that always uses DSG cars in Manual, so you can imagine how I’d feel about bike gears that change for me.
I’m shocked as to how much bicycles cost these days
although you needn’t spend too much to get something absolutely fine.
After a time though for every £1000 more you spend all you end up doing is shaving a few hundred grams off the weight - and also at the top end, get components that are just not as robust as their cheaper variants as sponsored bikes just don’t need parts that last as long as performance and weight are more important than longevity.
Most people probably could lose a single kilo in weight quite easily which would be the same in weight loss as spending an extra 3 or 4000 on a bike.
But spending a fortune on a bike you don’t need is probably in the same category as spending a fortune on a car which can do 400Km/h and driving it solely on the road.
There was a time during COVID when the waiting list for a bike was months.
Those days are long gone with many online shops gone - Wiggle and Chain Reaction collapsed and were bought by Sports Direct so now sell only budget gear and bikes.
Many bike manufacturers are struggling including big name like Trek, Kona and Cannondale.
I guess unless you are an aficionado, people new to cycling bought their during COVID lock-down and don’t need another for a few years.
Not really a use for AI but vaguely on topic - I was listening to a podcast and they said people were thinking of re-commissioning the Three-mile island nuclear power station due to excessive electricity demands in the US.
That will be EVs, I assumed but no, Microsoft want to get it re-commissioned and buy all the electricity to generate clean, no carbon energy to power AI data centres as they use an extraordinary amount of power.
Yes, in its current form, we’d need to build a lot more nuclear to power AI. However, I hope we transition to more efficient chips over the next few years that would cut down power consumption.
I was going to write that that’s inevitable according to Moore’s law but that’s more the processing power.
Using less steps to reach a conclusion will result in less power requirements. Won’t that happen automatically with AI as it gets cleverer and needs to ‘think’ less to solve a problem?
It’s not discussed much in media but Microsoft has a crazy plan: erasing their CO2 footprint since their foundation day.
That implies all current electricity must come from low carbon intensity sources such as wind, nuclear, or solar. That electricity is not exclusively for AI, it’s just to keep their offices and Azure running. So, turn off coal plants and revive nuclear. The final piece is capture CO2 from air to account for past emissions.
A bit of scifi, but if it works, the world might be a better place.
Or as the Wall Street journal put it, solar and wind and other intermittent generation is no good for this sort of thing as someone might ask chatGPT where they can get a pizza in the middle of the night.
I’m surprised someone at the WSJ is not familiar with electricity trading. It seems the woke are taking over and they forget good old capitalism exists.
Take MISO network in the US, spans from Louisiana to Manitoba (CA). Any power plant can deliver to any customer within the coverage of the transmission network. The Sun goes down, but the wind is always blowing somewhere. The terms of these contracts are really funny. If an electricity supplier cannot deliver at certain point in time, this supplier buys from someone else but the contract is fulfilled. In extreme cases gas turbines are powered up in a few minutes. But, there’s a certain level of overcapacity in wind generation that allows to meet demand at any time.
The current architecture is very power hungry. However, there are a few potential avenues to cut power. These are still in the research stages, but I’m sure the big players are investigating as the power savings are too big to ignore.
I used it to cancel a service in one minute. Previously, I’d write out a letter in English, use google translate to convert it to German and then fix up and send.
Now I can go to chatgpt and ask: write a letter in german to cancel this service with ref # 324. A great time saver and a help for people like me who hate to do such admin and always putting it off.
