HT Wired Wisdom: Have we handed the keys for AI tools to people with zero foresight?

There’s no other way of putting it. The world of tech is in absolute shambles. I can’t find a more diplomatic way to describe things as they’re unfolding. OpenAI’s board deciding to fire CEO Sam Altman, and then being forced to backtrack. Nothing falling for Sunbird’s false promises and firing off a shot at Apple claiming to have iMessage ready for Android. And how can we have the world going pear-shaped without some sort of contribution from Elon Musk and his companies. There’s an advertiser revolt underway at X because of Musk’s bad habit of ill-through posts (unlikely he’ll agree), and this isn’t the first time. Shambolic.

Also Read: In this kerfuffle, no one is sure where OpenAI, Microsoft or Sam Altman stand

Also Read: A new, diverse OpenAI board takes shape as Sam Altman returns

I will not attempt to write a closing paragraph on the OpenAI kerfuffle, as I call it. This Silicon Valley drama is far from over. The possible silver lining, just in case HBO is paying attention – this could be the inspiration for another season of the popular TV series? On a more serious note, what’s crystal clear is that the very folks who are responsible for artificial intelligence systems that are increasingly enveloping everything we do, do not seem to have any foresight or any hints of an innate ability to predict consequences of their actions. Considering what the OpenAI board did by pulling the trigger on Sam Altman and having the entire weight of their utter stupidity boomerang on them, is astonishing. And these are part of the group of people tasked with building super-advanced artificial general intelligence (AGI). A system that’ll be generally smarter than humans. As I had noted in my piece illustrating the partnerships OpenAI had assembled with Altman at the helm, anything that goes wrong with GPT or Dall-E or anything else they build impacts consumer-facing products and enterprise customised services by the likes of Microsoft, Snap, Expedia, Duolingo, and Slack – and with millions of users potentially interfacing with just these names, yet we’re only scratching the surface.

Sam Altman at the wheel or not, OpenAI’s tech knots the fortunes of many

It is too early for me to say anything. Whether Sam Altman will be back at OpenAI, whether he’ll accept Satya Nadella’s offer and join Microsoft. Or the often-mentioned start-up he might unveil. I’m still astonished, safe to say. For Microsoft, as for many others, the stakes in OpenAI’s wellbeing are higher than you may imagine. For once, OpenAI employees, who’ve made “OpenAI is nothing without its people” trend on X, are making the board hear them loud and clear. Either put the house back in order, or they’ll leave (and won’t be short of job offers either, any of them).

Following through on what he said, Elon Musk’s X has gone ahead and sued Media Matters, for a report (read that here and make of it what you will) that claimed X was placing ads from big-name clients including Apple, IBM, Oracle, Xfinity and Bravo (NBCUniversal) “next to pro-Nazi” content. It isn’t clear if that report had any bearing on advertising spends by any of the advertisers named in the report – though Yaccarino’s public statements do seem to suggest some have paused spending. Musk and X aren’t letting this go. Despite the bravado and inevitable allegations that the report is misleading or manipulating facts, the actual fact of the matter is a lot of what they’re saying is true. X’s reinstatement of previously blocked accounts for repeat offenders (Donald Trump included, Musk tried his best to make him an active user again) and the downward spiral of quality of conversation on the platform is pronounced. And for all of Linda Yaccarino’s attempts to put forward a brave face, Elon Musk goes ahead and does this . Feel free to say what you’re thinking.

MESSAGE

When the previous week began, no one could really have predicted tech making an absolute joke of themselves by the time the week was done. Apart from Silicon Valley and X feat. Elon Musk finding things going pear shaped in their own little worlds, a tech start-up bit off more than they could chew. Nothing, the folks behind the impressive Nothing Phone (2) and before that the Nothing Phone (1) decided to save humanity from the monopoly of Apple, and somehow bring iMessage functionality to Android phones. Their Android phones (actually, just the Phone 2 for now), with something called Nothing Chats. But wow, bringing Apple iMessage to Android, before Apple ever got round to considering it?

Nothing’s iMessage hack for Android is great, if you ignore the many red flags

I’d dug deeper into the specifics of how this would work, before the app rolled out for all users. That happened later in the week. Having a message sent though a Mac Mini farm, didn’t make for pretty reading. And then the persistent questions about encryption, promises of integrity of data and its subsequent storage. Therefore, no surprise Nothing Chats was pulled from the Play Store within hours of going Live. I’d guess the project has been shelved for now? Hang on, this is categorised as “several bugs” now? But I do not see a way back. At least not in this form. And confidence of anyone who intends to use it, is at its lowest ebb. There’s no way back after losing someone’s love and trust because of your own fault.

The reliance was on Sunbird’s service (though their own app has as yet not released; that should have been a red flag for Nothing’s decision makers). Yet, Nothing simply seemed to believe and go along with Sunbird’s promises that iMessages generated on a Nothing Phone (2), routed via an Apple Mac Mini in a server farm somewhere in the world and then delivered to your friend who has an iPhone, would be end-to-end encrypted. They weren’t. Plain text logging. Unencrypted data stored, including shared media.

Enthusiasm is fine. Unfortunately, the “Sorry Tim” thumbnail Nothing used to introduce Nothing Chats, hasn’t at all aged well. Perhaps, messaging services are meant to divide phone users. Those who’d prefer robust, encrypted system compared with a lax, bypass platform as the foundation. Think, hard.

LEAP

For years, we collectively embarked on a frenetic (and it was an extended period of that) activity in the search for a perfect smartphone experience. Whether we’re anywhere close to it or not, difficult to say. But there is now a concerted pitch for an alternative interface to phones. At least for some tasks. Are we as humans now so evolved (or made a quick dash to the opposite end of that scale, the less bright one) that even smartphones are cumbersome to use. Nevertheless if done smartly, the idea of blending AI’s foundation with focused wearables, may have considerable utility.

(Premium): Artificial intelligence, gadgets and the search for a hardware moonshot

Humane’s AI Pin, talk of the town for now, isn’t the first wearable that does focused AI well. But it may well be the most prominent, and therefore a benchmark. Also playing an important role in getting conversation going. This is no doubt equal parts bold and risky wearable, and at first glance, seems capable of a lot. Voice conversations with the ChatGPT bot, a compact projection system you can see on your palm for visual cues to your queries and the gamut of managing phone calls as well as health tracking. But what really is the big use case that’d justify you parting with $699 (around ?58,200; these ship in early 2024, and pre-orders began on November 16) for the AI Pin.

AI’s scope is widening. With that, every device that in any way plugs into that grid, is destined to become smarter. Even if not genuinely smarter, at least more capable. Apps on your phone, smartphone cameras, smart speakers, smart displays, and even your good old Windows PC with Copilot assistant, just some examples.

I had pointed out that any AI device, app or experience will only be as smart as the weakest point of the AI that underlies it. AI isn’t infallible. Not even by a long shot. Google couldn’t control the mistakes during its first demo of the Bard chatbot over the summer. It was asked, “What new discoveries from the James Webb Space Telescope can I tell my 9-year-old about?” The chatbot suggested it “took the very first pictures of a planet outside of our own solar system.” That, as it turns out, was done in 2004 by the European Southern Observatory’s Paranal Observatory’s NAOS+CONICA instrument. Not to rain on Google’s parade (they’ve done very good work since with AI and Bard in particular), but it just shows we cannot simply trust AI. That’ll percolate to AI powered gadgets too.

Not to say it won’t ever happen. In fact, it’ll be sooner than you imagine. AI is smartening up, eliminating the errors. Amazon, Google, Microsoft, Meta and OpenAI (whatever remains of it after the drama subsides) will drive the pursuit. The likes of Samsung, which has announced an AI call translator arriving next year, will complement the umbrella efforts with specific utility and focus.

Written and edited by Vishal Shanker Mathur. Produced by Md Shad Hasnain.