Tech Nuggets with the Technology

Sunday, November 10, 2019

The post-exponential era of AI and Moore’s Law

My MacBook Pro is three years old, and for the first time in my life, a three-year-old primary computer doesn’t feel like a crisis which must be resolved immediately. True, this is partly because I’m waiting for Apple to fix their keyboard debacle, and partly because I still cannot stomach the Touch Bar. But it is also because three years of performance growth ain’t what it used to be.

It is no exaggeration to say that Moore’s Law, the mindbogglingly relentless exponential growth in our world’s computing power, has been the most significant force in the world for the last fifty years. So its slow deceleration and/or demise are a big deal, and not just because the repercussions are now making their way into every home and every pocket.

We’ve all lived in hope that some other field would go exponential, giving us another, similar, era, of course. AI/machine learning was the great hope, especially the distant dream of a machine-learning feedback loop, AI improving AI at an exponential pace for decades. That now seems awfully unlikely.

In truth it always did. A couple of years ago I was talking to the CEO of an AI company who argued that AI progress was basically an S-curve, and we had already reached its top for sound processing, were nearing it for image and video, but were only halfway up the curve for text. No prize for guessing which one his company specialized in — but it seems to have been entirely correct.

Earlier this week OpenAI released an update to their analysis from last year regarding how the computing power used by AI¹ is increasing. The outcome? It “has been increasing exponentially with a 3.4-month doubling time (by comparison, Moore’s Law had a 2-year doubling period). Since 2012, this metric has grown by more than 300,000x (a 2-year doubling period would yield only a 7x increase).”

That’s … a lot of computing power to improve the state of the AI art, and it’s clear that this growth in compute cannot continue. Not “will not”; can not. Sadly, the exponential growth in the need for computing power to train AI has happened almost exactly contemporaneously with the diminishment of the exponential growth of Moore’s Law. Throwing more money at the problem won’t help — again, we’re talking about exponential rates of growth here, linear expense adjustments won’t move the needle.

The takeaway is that, even if we assume great efficiency breakthroughs and performance improvements to reduce the rate of doubling, AI progress seems to be increasingly compute-limited at a time when our collective growth in computing power is beginning to falter. Perhaps there’ll be some sort of breakthrough, but in the absence of one, it sounds a whole lot like we’re looking at AI/machine-learning progress leveling off, not long from now, and for the foreseeable future.

¹It measures “the largest AI training runs,” technically, but this seems trend-instructive.

Voi raises another $85M for its European e-scooter service

Voi Technology, the “micro-mobility” startup that operates an e-scooter service in a 38 cities across 10 European countries, has raised an $85 million in Series B funding.

Backing the round is a mixture of existing and new investors. They include Balderton Capital, Creandum, Project A, JME Ventures, Raine Ventures, Kreos Capital, Inbox Capital, Rider Global, and Black Ice Capital. The new funding brings the total raised by Voi to $136 million.

Eagled-eyed readers will have noticed that, based on our previous Voi coverage, the total figure is $32 million short. That’s because not all of Voi’s previous Series A commitment was cashed in after the company was offered more favourable terms for its $30 million Series A extension and therefore elected not to draw down the second tranche of its original Series A.

Launched in 2018, the company is best-known for its e-scooter rentals but now pitches itself as a micro-mobility provider, offering a number of different transport devices. These include various e-scooter and e-bike models, in a bid to become a broader transport operator helping to re-shape urban transport and wean people off using cars.

To date, Voi says it has 4 million registered users and has powered 14 million rides. More recently it has launched new, more robust hardware that has been designed to sustain the rigours of commercial e-bike sharing. The idea is that more suitable hardware will help e-scooter companies improve margins since more rides can be extracted from the life-span of each vehicle.

On that note, Voi says it will use the new funding to develop “strong profitable businesses” in the 38 cities where it is already operating, as well as increase its R&D spend to improve its technology platform and products. Earlier this year, the company announced that it is already profitable in the cities of Stockholm and Oslo.

“Clearly, we feel we are on track to achieve this in more of our cities and that is our aim,” Voi co-founder and CEO Fredrik Hjelm tells me. “At this point, a key focus for us is to ensure we continue to increase the lifetime of our e-scooters, forge key partnerships and continue to work in those cities which provide the best conditions for a profitable e-scooter business”.

Hjelm says that Voi’s version 2 scooters are projected to last over 18 months, which means the company should be in profit before it needs to raise again. However, he wouldn’t be drawn on when that might be.

With regards to R&D and improvements to the Voi platform, the company will continue to work on the lifetime of its e-scooters, in addition to improved repair management via integrating “predictive diagnostics”.

Hjelm also says Voi is developing “AI-powered” fleet management and more generally the platform’s capability to support future product portfolio expansion. In other words, we can expect new micro-mobility device categories in the future.