The volume of political news over the past few months has been deafening. Mainly coming out of the US, of course, but everywhere else too. Trying to make sense of it has been a little overwhelming, even for those who make a living doing so.
So it is with some trepidation that I wade into the other big news story that has been pushed off the front pages, which is the blistering (and almost daily) advances of AI and the inevitable emergence of geopolitics around it. And (almost inevitably) a vicious personal vendetta between two of the richest and most powerful AI titans in the world.
Let’s start there, because that is the sizzling one. I refer, of course, to the spat (an understatement) between OpenAI’s Sam Altman and xAI’s Elon Musk. This has become a blood feud of sorts, harking way back to a time (2018) when OpenAI was small and anonymous, and Musk owned part of it.
Musk wanted Tesla to take it over, because he needed AI tech to be the new fuel for Tesla’s self-driving aspirations. OpenAI’s board demurred, and Musk left in a huff, apparently claiming that it could never succeed without him. It has escalated from there. Which might have been an amusing spectacle, except that these guys have their hands on some of the most powerful technologies ever invented. So having the combatants throwing in a soupçon of spite and malice into their plans may not be that good for the rest of us.
In any event, all of this may have disappeared into forgotten history, but in 2019 OpenAI started making moves to migrate from an altruistic non-profit venture to a profit-oriented business. This requires a new personality — sharp elbows and a take-no-prisoners attitude. The move to profit pissed off Musk no end, presumably because he wanted that space for himself.
Bullish
Meanwhile, he had gone on-line media shopping, famously scooping up Twitter in October 2022 for $44 billion. Still bullish on AI and now faced with the growing juggernaut of OpenAI, Musk offered to buy it in 2025 for $94.7m, opening a door for this snarky response from Altman — “no thank you but we will buy Twitter for $9.74 billion if you want.”
Ouch.
The legal side had gotten nasty too. In 2024, Musk filed a lawsuit against OpenAI and Altman, alleging that the company had violated antitrust laws and strayed from its founding principles. Musk sought to block OpenAI’s conversion to a for-profit entity, but a federal judge denied his request for an injunction (this is an ongoing matter). Altman countersued, accusing Musk of “bad-faith tactics” designed to hinder OpenAI’s operations and gain control of advanced AI technologies for his own benefit.
More public barbs were lobbied. Altman accused Musk of being “personally insecure,” and “just embarrassing to watch.” Musk accused Altman of being a “swindler” and has called him “Scam Altman.”
All very schoolyard stuff.
Social media rival
In any event, the latest episode has just arrived. Multiple rumours have OpenAI planning to launch a social media competitor to X, based on OpenAI’s tech. Details are not clear yet (apparently AI image generation will play a big part), but odds are that this will happen. Why do this? Because OpenAI would want to use the data generated by social media users to better train their AI models.
Is Altman doing this just to annoy Musk? Probably not, but it’s an enticing thought.
Which brings us to the other news in AI.
There are many contenders for the crown in this business, if there is indeed only one crown. OpenAI is the current monarch, but the court is full of pretenders, aspirants and jealous cousins, including Anthropic, Meta, Alphabet, Microsoft, Nvidia and Amazon in the US, Mistral in France, and DeepSeek and Baidu’s ERNIE in China.
The rate of innovation within these companies beggars belief. New AI models are released almost every week. Not just minor upgrades, but startling new sets of capabilities across text, image, and video and increasingly all three simultaneously.
Small powerful AIs to fit on non-connected devices, AIs that can ‘reason’ across multiple domains and tell you how they come up with answers, AIs that can send agents out to the world to complete disparate and parallel tasks before consolidating and returning to the user.
And then lower down the AI model stack — models that can write software code at genius level, models that can do math at PhD level, models that can diagnose disease, that can discover new science, that can help build synthetic biology. And, tantalizingly, AI models that can build better AI models.
Tsunami
Trying to keep up with this tsunami of announcements are the ‘benchmarks’ — those carefully constructed third-party tests which try to keep track, of which AI is superior to its competitors across several axes. They too are being swamped into almost instantaneous obsolescence as new releases and versions pour out of the industry firehose.
Any commentator looking to provide careful analysis of the changing landscape — what is out there, what we should be using, what it should be used for, where we should place our bets — has a near impossible task. We are in an age of unprecedented confusion — a Cambrian chaos of new creatures evolving and metamorphosing in real-time.
For those who try to keep it all in focus, there is a further problem to contemplate. It is whether each successive announcement is an indication that AI is getting ‘better’ or at least getting better at the same rate as before. It is called the scaling problem, and the top AI scientists spend a lot of time getting annoyed with each other’s differing views on this subject. Mainly because one view means we humans are still OK for a good long while, the other means the machines are taking over from us soon.
Which all boils down to this: AI is going to change the world. When? I guess you’ll know it when you see it. Like when you lose your job for which you studied 8 years to acquire. Or your incurable cancer suddenly gets a cure.
Finally, the geopolitics. The Chinese are here, as of the release of DeepSeek R1 in January 2025. Their big frontier AI models are shockingly good. Why shockingly? Because no one in the West believed that they could possibly compete in this most lofty of intellectual pursuits. Of course, they could and did.
There are few global strategists who would argue that he who builds the smartest AI will be the next superpower, possibly forever.
The last great arms race is under way, and its stakes are, well, the biggest.
[Image: TechCrunch – TechCrunch Disrupt San Francisco 2019 – Day 2, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=92008259 and https://www.flickr.com/photos/tedconference/33944890310]
The views of the writer are not necessarily the views of the Daily Friend or the IRR.
If you like what you have just read, support the Daily Friend