CES 2023 Day 1 – Foot and mouth

CES 2023 – Almost normal

• Despite a lot of hype, the footfall at CES 2023 is below pre-pandemic levels.
• There are a number of mitigating factors:
  • First, China, which is largely absent outside of the really big players like TCL who have a local presence.
  • Despite geopolitical tensions, China remains a crucial part of the technology industry and it is the current wave of Covid that is keeping it at home.
  • Second, economy: Everyone is belt-tightening which in many cases has meant a lower representation at the show as well as fewer visitors overall.
  • Valuations: for early-stage companies and others that are pre-revenue or loss-making have been eviscerated in the last 12 months.
  • This makes it much harder and highly dilutive to raise money and so much closer attention is being paid to expenses particularly related to attending shows such as this one.
• Even with these two factors, I had expected that pent-up demand to visit the show would have been higher.
• All areas are quieter than normal.
• This is to be expected at the convention centre where the new West Hall is in use for the first time which has enabled the show to spread out somewhat and be less dense.
• However, the Venetian / Sands Expo is exactly the same and here too, there is more space, shorter queues and one can actually walk around without fighting one’s way through a crowd.
• This is why I suspect the CTA (which runs CES) chose to have two of the days on a weekend as it knows that this will concentrate the numbers into 2 days rather than 4 and ensure that the venues on at least 2 days are pretty busy.
• Any quietness on the other two days can be put down to the weekend which is exactly what I expect to happen.
• That being said, I would estimate that the show is running at around 80% – 90% of pre-Covid years which is more than enough to call it a success.
• I suspect that this positive momentum will ensure higher attendance in 2024 by which time I am hopeful that China will return to the show.

Automotive – talking but not thinking.

• Gone are the aspirations of cars that drive themselves and back is the idea that in 2025, the driver will be able to have a conversation with the vehicle which in my opinion is equally fantastical.
• Many OEMs, Tier 1s and other suppliers are once again touting voice as an easy and fun way to interact with one’s vehicle, but I remain unconvinced that the AI has evolved enough to allow this to work much better than it did when the idea was first touted 6 years ago.
• 6 years ago visitors wandering the North Hall would be inundated with promotional videos of vehicle digital assistants assessing the mood of their humans and suggesting to them how they should drive the vehicle and what music they should listen to.
• However, the realities of AI got in the way of this as users started to realise that beyond playing music and setting timers, Alexa, Google Assistant, Siri and so on were not of much use.
• However, the advent of ChatGPT seems to have changed this attitude again in another move that I think is misguided.
• ChatGPT is very good at producing text that gives the impression that the system knows what it is talking about but the reality is quite different.
• For example, getting the system to contradict itself is quite a simple task as is getting the system to express political views from all ends of the spectrum despite OpenAIs best efforts to ensure neutrality.
• The popularity of ChatGPT and the buzz it has created is being put to good use by OpenAI which appears to be seeking to raise money at a crazy valuation of $30bn.
• In this climate, this will be quite a feat as the company has no revenues that I am aware of.
• This “buzz” has bled over into CES where it is now being assumed that ChatGPT heralds a new era in human-to-machine communication but I do not think this is likely.
• One reason for this is that in order to control the dataset, ChatGPT is frozen in time meaning that its knowledge does not evolve.
• For a chatbot, this is pretty useless and so I think much more work is needed before humans and machines can hold meaningful conversations.
• This has not stopped the industry from jumping back on the bandwagon and hyping up voice once again now that autonomous driving has so badly failed to live up to its promises.
• I suspect that this new optimism will also fail to deliver as while ChatGPT is quite impressive, it has not solved any of the issues that led to voice assistants failing to live up to expectations.
• I have long liked the idea of voice as the user interface for the vehicle, but the problem is that the AI to deliver it effectively remains miles adrift of where it needs to be.
• However, it makes a good story for those that have forgotten the hype cycle of voice 6 years ago.