Bold AI predictions coming true (and it's only March)

(covering the week of March 23, 2026)

One of the best questions I get is ‘how do you manage to digest all this AI news?’. I say ‘best’ because it’s usually followed by ‘how are you not depressed???’

Well, despite the wide range of sources and stories, I only trust a few voices in this space. I rely on them to share the news, stories, topics to follow and then go dig. One of these voices is Casey Newton who puts out Platformer three times a week.

To prove how great he is at this, check out one of his bold 2026 predictions:

OpenAI retires Sora. A year from now, I suspect that we will look back on Sora as a moment when OpenAI was at its most unfocused. At a time when Google and Anthropic were wowing customers with advancements in AI code, OpenAI was launching a TikTok-like social network that immediately came under fire for its lack of copyright guardrails and vulgar embrace of brainrot. While it was a hit on release, Sora usage has tapered off sharply since then. Assuming Sam Altman is serious about diverting more resources to ChatGPT, he’ll wind down Sora development early next year and bring those creative tools and social features to the place where he’s already winning: ChatGPT.

Now, check out this week’s first story.

Let’s get to it.

8ish stories – Here we go.

Story 1: OpenAI retires Sora (Reuters)

BREAKING this morningAnthropic admits it’s testing a new Mythos model – Capybara – that would be its most powerful model yet. (Techzine)

We’ll keep an eye on this.

Story 2: Wordpress is letting AI agents make websites. (TechCrunch)

Is this the future of web design, development and management? Sure looks like it. Whatever you want…just ask the agent to make it. Write it. Update it or maintain it.

Are we surprised?

Story 3: Surprise! Google is rewriting headlines with AI. And there’s nothing you can do about it. (9to5Google)

You know that AI summary you get that replaced AI overviews that replaced blue links when you search for something? Well now Google is shortening the headlines, condensing the info and combining all of it into a more streamlined summary, potentially altering how the article reads or is positioned.

At what point do we just give up and read what you want us to read, Google? I mean, what’s the point here? Why are you doing this?

Something to track.

Story 4: Advertisers can’t prove ChatGPT ads work. (Search Engine Land)

(from the article) “If you’re considering ChatGPT as an ad channel, the lack of performance data means you’re spending blind — with no reliable way to prove ROI to clients or stakeholders.”

Story 5: “Guardrails”

What happens when 93% of prompts are approved without reviewing? Claude says ‘well fine, here’s auto mode for Claude Code.’ (Shelly Palmer)

And that’s what we have now. Claude Code just does it.

Human in the loop, eh?

BREAKING last night Federal judge temporarily blocks Pentagon from branding Anthropic a security risk. (AP)

Story 6: The age of authorship (Good Thinking)

This came up in our AI Council discussions…and it’s a pretty damn good assessment of the state we’re in when it comes to how and why we should be using AI.

Short answer: Because we should. While we can.

Story 7: Remember when we said AI was an intern?

Yea, we’re not there anymore. Here’s an article in HBR about scaling AI:

To scale AI agents successfully, think of them like team members.

Story 8: What’s missing from AI outputs? Pride. (Christopher Penn)

Amen.

3 Nopes.

1. AI is blamed for another CEO stepping down. (Gizmodo)

2. Sycophantic AI can undermine human judgement. (ars Technica)

3. Number of AI chatbots ignoring human instructions is increasing. (The Guardian)

-Ben