AI too powerful for the public, already in the office.

Hello,

This is Simon with the latest edition of The Weekly. In these updates, I share key AI related stories from this week's news, list upcoming events, and share any longer form articles posted on the website.

This week, Anthropic released a new model called Claude Fable 5. It's a variation on the much-talked-about Mythos model, which many have deemed too powerful for the public. To make Fable suitable for general release, Anthropic has built in a number of guardrails — including handing sensitive topics such as cybersecurity, biology and chemistry over to the older Opus 4.8 model to minimise potential misuse.

What I found interesting was that Fable 5 was already available in my Enterprise edition of Claude this morning. Given how powerful it's meant to be, I'd have thought our IT team would want to "test" it first before rolling it out across the business.

And it's not just powerful — it's also more expensive. I asked Claude to build a presentation slide this afternoon: a task that took six to eight minutes and, according to my usage dashboard, cost around $12 in tokens. It sparked an interesting conversation in the team. We're all keen to try the new model and get a feel for what it can do and how it might help with our jobs — but we're also very mindful of what it's costing the business.

I've written recently about being aware of what you're spending on AI, and whether it's actually delivering a return. With a model like Fable 5 now in employees' hands, that question becomes even more important.

Right now, I don't think many companies are asking their staff to think about the value of their AI work. On the whole, they just want people using the tools they've been given, with the aim of becoming more efficient and getting more done in less time. But I genuinely believe a lot of businesses are in line for a huge shock over the coming months — when they start adding up their AI bills and comparing them to the output those tools actually deliver.

To give you a sense of scale: in my Enterprise Claude account, I appear to have a spend limit of $15,000 a month. On the one hand, it's amazing to have that much room to use AI in my role. But on the other, if I get anywhere close to it, I can imagine they’ll be a conversation to be had.

So I'll leave you with this. A model that's been described as too powerful for public release just landed on my desk at work. Does that excite you, or does it make you feel terrified?

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

In this section, I’m going to bring to you a real world example of AI use. This week we look at how Morgan Stanley achieved productivity gains with a debrief tool.

OpenAI has filed confidentially with the US Securities and Exchange Commission for a public listing, according to Bloomberg and multiple other outlets. The company is working with Goldman Sachs and Morgan Stanley on a potential stock market debut as early as September, targeting a valuation of between $730 billion and $850 billion — with some analysts predicting it could exceed $1 trillion by the time it lists. OpenAI has raised more than $180 billion from private investors and continues to burn through cash at a significant rate to fund model development and infrastructure.

Why it matters: An OpenAI IPO would be among the largest technology listings in history, cementing AI as a distinct investment category rather than a feature of existing tech. For business leaders, it is a marker of how quickly the competitive landscape is shifting — and a signal that the window for organisations still evaluating AI strategy is narrowing.

At its Build 2026 developer conference, Microsoft unveiled seven proprietary AI models under the MAI brand, including a reasoning model and a coding model designed to compete directly with offerings from OpenAI and Anthropic. The flagship coding model, MAI-Code-1-Flash, achieves 51% on SWE-Bench Pro despite running at a fraction of the cost of larger models. Microsoft said its enterprise focus is firmly on "governance, security, and model choice" — and introduced a feature called Frontier Tuning, which lets companies fine-tune model behaviour using their own data without that data leaving their compliance boundary.

Why it matters: Microsoft moving to develop its own models is a significant strategic shift — and a warning shot to OpenAI, in which it has invested heavily. For enterprise customers, the more relevant development is Frontier Tuning: the ability to adapt model behaviour inside a company's own governance boundary has been a blocker for AI adoption in healthcare, financial services, and regulated industries. That constraint is beginning to ease.

A PwC study of 1,217 senior executives across 25 sectors found that 74% of the economic value being generated by AI is concentrated in just one-fifth of organisations. The leading companies are not simply deploying more tools — they are using AI to pursue new revenue and reinvent business models, and are 2.6 times more likely to say AI is enabling them to change how they operate fundamentally. The gap between leaders and the rest is widening: top performers generate 7.2 times more value from AI than their peers and report profit margins four percentage points higher.

Why it matters: For the majority of organisations still running pilots, the study is a blunt diagnosis: activity is not the same as value capture. The companies pulling ahead are making different choices — embedding AI into strategy and growth, not just efficiency — and the gap is becoming structural.

Thanks for reading, and see you next Thursday.

Simon,

Was this email shared with you? If so subscribe here to get your own edition every Thursday.

Enjoying Plain AI? Share it and get a free gift!

If you find this newsletter useful, why not share it with a couple of friends or colleagues who would also benefit? As a special thank you, when two people subscribe using your unique referral link, I’ll send you my "Exclusive Guide: Supercharge Your Work with NotebookLM." It’s a practical, no-nonsense guide to help you turn information overload into your secret weapon.

You can also follow me on social media: