Long story short, MBrownlee's review literally killed that AI company's stock.
The funny thing is that some EXTREME reactions came online from AI "bros" as well as AI corpo bosses, who all blamed Brownlee for very negatively review a very negative product.
I'll hazard a bar table psychoanalysis: most of these people know that their product is just as bad as the Humane AI one. They know that they're hacky, poor quality products with no actual appeal except through AI Hype.
I think the reactions, and virulence of the reactions, are all telling of people who know their company is going under within the year. And obvious scammers of course. Massive amount of "salespeople" complaining.
That company's stock was always going to shit the bed as their product was always going to fail to live up to the hype. I honestly don't know why the media was fluffing them and the other company with a similar (but also distinctly different) device back at CES at the start of the year. Some pointed out it had significant shortcomings (that almost certainly would not be fixed much if at all by the time of release), but overall still fluffed this stuff, it had like 2005 era Engadget vibe (where they just act like all tech is magic). But like I pointed out, there's a bunch of AI evangelists that are full blown cult mentality, so not surprised at the backlash.
There's some podcasts I've been listening to recently that make the case that we might actually be close to peak AI. Even the AI companies admit that in order to take the next step in training will require an order of magnitude increase in computing power, and they don't actually have the money to get there since none of them are actually making money compared to how much they're spending (and many of them do not have the capital to get there). That's why Sam Altman has been trying to push to monetize AI so much, and almost certainly why the board booted his ass last year, as that was somewhat explicitly not supposed to be what he was doing (its more the overall manner he was going about doing that which he was not supposed to be doing per the non-profit's clearly laid out intentions; like massively overhyping AI and him more or less becoming intrinsincly linked to Microsoft, as well as him starting up other AI companies). Further, OpenAI canned their more efficient fork that Microsoft was demanding (because AI is so resource hungry even for the simple chat prompt shit) because it wasn't more efficient at all.
But it highlights why all of a sudden companies are now pushing to run LLMs on individual devices, as its an area they can increase (since for instance huge disparity between the bit for AI on SoCs compared to even GPUs let alone the AI focused stuff now being built). Its another race for companies to push newer faster hardware to make people upgrade. Plus it offsets things to the consumer (cost of development, power use, network use). The training is the most resource intensive part and that will still be done on their side, but with how much they want to push AI it'll clog data centers up. Plus they likely also have to for the European privacy stuff (with the US starting to somewhat follow suit - California and even at the federal level there's some push for privacy rights, etc), and also this way it'll work with content on the device (which helps limit their liability as far as copyright/etc; its kinda like how companies ignored the legality of the music people had on their computers letting people use that in their library while pushing them to legally buy new music; and almost guaranteed they'll push people to subscribe to license content libraries to use or lock-in to companies that were - saw a bit of this with Giphy). Its basically a stopgap fix so they can keep hyping the hell out of AI.
Also listened to some older episodes that were about crypto, and the similarities between the 2 were really highlighted, and especially how delusional people are about them. It really is like tech industry upset they lost out on the crypto hype train money so decided to fluff AI for the same (seemingly intent on ignoring the rampant fraud that occurred/occurring in crypto). I'd argue AI has bigger risks and bigger blind spots too, so its even more alarming at how the Tech industry is pushing it so hard. The situation in the UK where postal employees were sent to prison because of a known software glitch should raise alarm bells.
Which, if people are interested in the podcasts, there's some specific episodes of Behind the Bastards (I linked before about AI leading to cult like mentality in tech industry; was an ok series on Steve Jobs recently as well; had some decent ones about SBF and how full of shit his "philanthropy" was), but Better Offline, Search Engine (Crypto Island was a nice condensed view of things, and had some other solid episodes but its often not about Tech industry at all), and Tech Won't Save Us are ones that ditch the Tech hype train (and even focus on how we transitioned from the almost overwhelmingly positive view of Tech circa early 2000s-2010s to where we are now). Mozilla also has (had? hasn't been updated in several months) but I haven't listened to it yet so not sure if its any good (episodes seem short so probably just simple overview of topics).