MWC 2024 Day 3 – The AI effect

AI disguises the also-rans and pushes ORAN to the periphery of consciousness.

Start-ups – AI pollution

• The best part of every trade show be it CES, MWC or similar is the opportunity to descend into the bowels of the show and filter through thousands of start-ups which have a huge range in terms of quality.
• The problem this year is that 70% of the start-ups have put AI somewhere in their name or their one-sentence description meaning that the obviously mediocre offerings have become much more difficult to identify.
• To fix this problem, I ignored every company that used the AI term and focused on those that did not and here I found the quality to be far better.
• I suspect that this is because to get noticed without the AI moniker one has to have something that stands on its own merit.
• One theme that I found particularly interesting is gestures as an input method which Apple has brought back into the limelight with the release of the Apple Vision Pro.
• Two companies in particular stood out.
  • First, Doublepoint: which is based in Finland and uses the sensors available on Android Wear devices to enable pinch to click as well as flick to scroll.
  • One can also use hand movements to act as a mouse on an iPad with a pinch as the mouse click.
  • The secret sauce here is error detection as Doublepoint can tell when fingers are pinched deliberately as opposed to when they are not.
  • Apple got around this problem by requiring a double pinch which almost never happens by accident.
  • Android Wear is a bit limiting in terms of what it allows the company to do but with its prototype device on one’s wrist, one can point at lights and turn them on and off by pinching which is a pointless, but very cool gimmick.
  • It is worth remembering that pointless gimmicks often cause devices to sell in their millions.
  • Second, Neuranics which is a company that uses magnetic fluctuations in the body to read body signs and it claims this can also be used to decode gestures.
  • The company is currently demonstrating a prototype device that can work anywhere on the left side of the chest which can very accurately (presumably to medical grade) provide a full ECG.
  • The company says that this is good enough to get FDA clearance, but I see this more as a proof of concept of what the company can do as medical devices is not its goal.
  • A device on the wrist will be able to measure heart rate and HRV which offers an alternative to light which can struggle to work well on people with darker skin.
  • Most interestingly the company thinks that this could also be used for gesture-based input which could be both very accurate and detailed.
• This is a long-term theme as Apple has demonstrated that when it is implemented well, it is a perfectly good system for input although typing is an issue.
• Using these solutions rather than very expensive sensors offers a way to get pricing down in terms of sensors, power consumption and weight.
• It is an emerging theme which has a new lease on life and one worth keeping an eye on.

ORAN – ORAN? What ORAN?

• Missing in action at this year’s show is OpenRAN (ORAN) which was all the rage last year but has been completely sidelined by AI in 2024.
• ORAN is the idea of having all of the interfaces between different pieces of network equipment open and available meaning that operators can pick and choose different vendors for each piece.
• I was expecting this to be a big deal this year as for years ORAN has suffered from poor performance but with a recent update, this problem was largely solved.
• Hence operators should be going gangbusters on mixing and matching equipment and getting excellent prices, but at the show this year, ORAN is hardly anywhere.
• No one is talking about it, there are no banners anywhere leaving it to the existing ORAN supporters to wave the flag.
• It seems to me that AT&T’s deal with Ericsson should have pushed this to the top of the agenda but the nature of this deal has done precisely the opposite.
• This is because instead of picking and choosing components from different vendors, AT&T has taken the whole lot from Ericsson making this look just like any other deal for network infrastructure.
• In my opinion, this really undermines the whole proposition of ORAN as it makes it look just like any other network deal that uses closed interfaces.
• The only real winners here are 3^rd parties who offer network software as they will still be able to plug into the network even if the whole lot has come from one vendor.
• AT&T could theoretically use 3^rd parties for capacity upgrades but there is no indication that it will do this, and I suspect that Ericsson will do everything it can to ensure that it keeps the business.
• This means that ORAN could still take a piece of the market, but it does not look like the equipment vendors that specialise in the mix-and-match idea will get much of a look in.
• Hence, it looks like tough times will continue for them despite the large boost in confidence that ORAN has just received.