Today brought so many stories worth “notes and observations” that I thought I’d try something new – a flash newsletter of sorts, with commentary on stories that pushed my eyebrows up a bit more than usual. Each of these items is worth a full-fledged long-form piece, but it’s Friday, so let’s be brief:
Do You Trust OpenAI? Over the past two years I’ve been warning that when it comes to long promised “user agents” that work on our behalf, AI companies would inevitably adopt the big tech playbook of providing centralized services that they control, ensuring that consumers are dependent on their platforms and by extension, locked into their services. This architecture is anathema to true innovation in a modern data economy, but inevitable given the capital constraints of current AI models. Well, this morning brought news of OpenAI’s “Agent,” which purports to “take over” our computers and take action on our behalf. As I’ve asked, over and over, is this the way we want the future to unfold? Who exactly do we think OpenAI’s agent really works for? Hint: It’s not us, anymore than Facebook, Amazon, or Google ended up working for us.
The Death of Woke in AI = State Sponsored Censorship. Related: Today also brought the news that the United States Government is demanding that any AI models it contracts with (and that’d be ALL OF THEM) must prove they are not “woke,” whatever the f*ck that means. And that’s kind of the point: We’ll give your industry billions of dollars (literally), but only if your product parrots MAGA talking points. This is … beyond fucked up. It’s government censorship, and a massive onramp to autocracy. Will the AI industry balk? Don’t bet on it.
Big Tech Pays For Silence. On the topic of Big Tech, and specifically Facebook, the company settled its long running lawsuit over its relationship with Cambridge Analytica. Some have noticed the interesting timing of the settlement: Right before the case would have required potentially embarrassing (or worse) public testimony from Facebook’s executives and board, which includes some of the most powerful actors in tech, including Marc Andreessen and Peter Thiel, both of whom have become major power brokers in Trump’s version of America.
And It Pays For Graft. This isn’t unexpected, but…you do know who the largest holders of crypto currency are, right? Yep, they’re the same folks who just dodged accountability in the Facebook trial, along with a long list of grifters who’ve used crypto to buy off our current administration, starting with the President. Today brings news that they’re getting what they paid for, legislation that will bring cryptocurrency into the mainstream of American economic life, including 401ks. What could possibly go wrong?
Oh Man, I’ve Seen This Movie Before. Lastly, Substack, which is perhaps the technology press’s favorite company to cover, raised $100 million at a “unicorn” valuation of $1.1 billion. The past few months have seen a spate of “Substack’s business model will never work” stories, and I have to say I find myself nodding along with them. Some of you may recall I founded and ran a company called Federated Media, which was essentially a Substack-like platform for bloggers, back when “I have a blog” was the equivalent of today’s “I have a Substack.” The big news in Substack’s funding round wasn’t the money – they’ve been trying to raise for three years now. No, the truly interesting bit was how the news was accompanied by a “softening” towards advertising as a revenue stream and “social media” as a strategy going forward. This. Will. Not. End. Well.
Happy Friday! Now go get that drink.
—
3 thoughts on “If You Trust AI, You’re Asleep. (At Least You’re Not “Woke”)”