The AI world is so full of hype, I wouldn’t blame anyone for shrugging at all the drama around the release and subsequent pullback of Anthropic’s Fable 5 model. After all, the industry releases new models all the time, and considering all the headlines from the year—which included the Elon Musk vs. Sam Altman trial as well as Anthropic’s dust-up with the Department of War—the Fable story might seem like just another day in AI.
But this is different. This is really the first time the government has stepped in to regulate a specific model release on the grounds that its capabilities could pose a national security risk in the wrong hands. However the events play out going forward, the incident marks a turning point: AI models have reached a level of capability where the government sees the need to take an active role in managing them. That sets a precedent that every user of AI should care about, since it means the level of intelligence available to you won’t just be a factor of cost but also what’s allowed. It may also depend on where you live, what data rules you accept, and whether a government or vendor leaves the model online.
The story of Fable 5
If you don’t follow AI closely, a quick recap: Fable 5 is the first generally available model in what the company is calling its “Mythos-class” models, a tier above Opus that Anthropic says has crossed a meaningful risk threshold in cybersecurity and biology. Fable 5 is the guardrailed version of Mythos that is supposedly safe for general users. It uses the same underlying model as Mythos 5, with additional safeguards that can block or downgrade certain cyber, biology, chemistry, and model-development requests. It also jumps Anthropic’s core model number, signaling a generational step forward from Opus 4.8, Sonnet 4.6, and Haiku 4.5.
{“blockType”:”mv-promo-block”,”data”:{“imageDesktopUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/03\/media-copilot.png”,”imageMobileUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/03\/fe289316-bc4f-44ef-96bf-148b3d8578c1_1440x1440.png”,”eyebrow”:””,”headline”:”\u003Cstrong\u003ESubscribe to \u003Cem\u003EThe Media Copilot\u003C\/em\u003E\u003C\/strong\u003E”,”dek”:”Want more about how AI is changing media? Never miss an update from Pete Pachal by signing up for \u003Cem\u003EThe Media Copilot\u003C\/em\u003E. To learn more visit \u003Ca href=\u0022https:\/\/mediacopilot.substack.com\/\u0022\u003Emediacopilot.substack.com\u003C\/a\u003E”,”subhed”:””,”description”:””,”ctaText”:”SIGN UP”,”ctaUrl”:”https:\/\/mediacopilot.substack.com\/”,”theme”:{“bg”:”#f5f5f5″,”text”:”#000000″,”eyebrow”:”#9aa2aa”,”subhed”:”#ffffff”,”buttonBg”:”#000000″,”buttonHoverBg”:”#3b3f46″,”buttonText”:”#ffffff”},”imageDesktopId”:91453847,”imageMobileId”:91453848,”shareable”:false,”slug”:””,”wpCssClasses”:””}}
On June 12, three days after the release, the government ordered Anthropic to block Fable 5 and Mythos 5 from every foreign national, including foreign-national employees working inside the United States. Anthropic said it could not reliably enforce that distinction and disabled both models globally. The government’s concern reportedly centered on a suspected jailbreak that bypassed Fable’s cybersecurity guardrails. Anthropic disputed the severity of the finding, saying the demonstration uncovered only minor, previously known vulnerabilities that other public models could also identify.
As I write this, that disagreement is still playing out. Cybersecurity leaders have urged the government to reverse the order, arguing that defenders need access to the same capabilities and that comparable tools are already available from American and Chinese competitors. Anthropic is trying to restore Fable, and models from other labs will almost certainly quickly rise to a similar level.
But even if this specific dispute is resolved, the precedent remains. A model can be released, integrated into workflows, and then disappear because a government draws a line around who may use it. For anyone building around a single model or vendor (and “building” might simply be leveraging it in crucial, strategic use cases), availability is now part of the risk calculation.
The capability threshold
Based on the early impressions from when Fable was generally available, the model delivers on its promise of a next-gen model. Despite ongoing controversies over how Anthropic chose to limit how Fable 5 deals with queries the company deems risky (more on that in a minute), users are seeing the power of the model. Fable 5 is designed for agentic work, meaning it can work autonomously on tasks for a long time—sometimes hours or even days—without losing context. The best way to use it, many say, is not to ask it to perform straightforward in-and-out tasks like writing an essay or telling you the best parts of a lengthy report, but to give it broader goals about what you’re trying to achieve, let it build the plan, then execute—however long it takes.
That brief window matters because it showed that this level of intelligence is no longer theoretical. The model was pulled back, but the capability threshold remains crossed.
A big part of what makes Fable so effective is self-correction. If you’re a regular user of Anthropic’s models, you’ll notice there’s no “Thinking” mode for Fable 5. That’s because adaptive thinking is always on: the model decides when and how much to reason on every request, and at higher effort levels it can reflect on and validate its own work. In practice, that means the things it’s assigned to do can be put on “loops.” As it works to achieve the goal, it can try things, evaluate the results, change course as needed, and try again. And it can do so autonomously.
In practice, this means that you can get much more ambitious with the kinds of tasks you delegate to AI. Instead of, say, assigning it to design a specific email campaign, or help format your newsletter, you can zoom out and tell Fable 5 to conceive and build an entire marketing strategy around your newsletter. That might involve reformatting your templates, building new landing pages, adjusting the publishing schedule, building a social campaign, and more. Theoretically, with the right access, it could then build all of that for you. All you’d need to do is evaluate its work. Over time, that evaluation would shift from happening throughout to mostly happening at the end.
That is the promise, anyway. The danger is that organizations may begin designing around that promise before access, cost, and governance are stable enough to support it.
This is the barrier that Fable 5 pierces (and certainly very soon, models like it): true agency. Right now, working with agents, while powerful, involves a lot of management: ensuring the plan the agent builds is correct, clearing up barriers that it encounters as it performs the task, and then guiding it to the best output—usually through multiple iterations on the task itself. In theory, a powerful enough model should be able to iterate on its own output.
That is what makes Fable’s disappearance more than a product hiccup. For a few days, users could test a different relationship with AI, then the capability vanished. We may have crossed the threshold technically without crossing it operationally.
The walls around frontier intelligence
TL;DR: The next class of frontier models—of which Fable 5 is the vanguard—is potentially poised to change the nature of how we use AI, if not work itself. However, using Fable 5 to its full potential was never just a matter of selecting it in your model picker or calling the API and letting it cook. The pullback made the more basic problem obvious: sometimes the model may not be available at all. I see three distinct barriers standing in the way:
1. Access and context: For an organization to successfully use Fable 5 to its full potential, however, it would require a large amount of access to the right context (the org’s information and data). Here’s where Fable 5’s power gets in its own way: Because Anthropic fears the model could be misused, it requires prompts and outputs from Mythos-class models to be retained for at least 30 days for safety monitoring, including in enterprise environments that would otherwise use zero data retention. Anthropic says the data will not be used to train models and that, on some third-party platforms, it remains inside the customer’s cloud environment. But companies cannot use Fable 5 under a true zero-retention arrangement.
This, along with some restricted categories where Fable 5 will throttle its power down to Opus 4.8, has sparked controversy among many power users, and it could limit Fable 5’s adoption with enterprise customers, since many companies will be reluctant to cede control over how their own data is retained and reviewed. Microsoft reportedly limited employee access while its legal teams assessed the implications for confidential and customer data.
The shutdown adds another layer. Even if a company accepts the privacy terms, secures the integrations, and builds the right internal controls, the model can still disappear because of a government order or vendor decision. Serious agentic systems will need fallback models, portability across vendors, and a plan for what happens when the most capable model is suddenly unavailable.
2. Compute: As a powerful model, Fable 5 is pricey to run. Anthropic priced it at $10 per million input tokens and $50 per million output tokens, twice the price of Opus 4.8. As I’ve written previously, one of the consequences of the agent era is that everyone is feeling the cost of compute, and with AI becoming a political wedge issue, everyone is likely to be compute-constrained for the next several months, if not years.
That does not automatically make Fable uneconomic. Some early users argued that it could solve hard tasks in fewer turns than weaker models, potentially lowering the total cost of completing the work. But that only makes task selection more important: the value depends on assigning work that merits the premium.
While Fable 5 and models like it seem destined to be the ultimate orchestrators in a company’s AI hierarchy, in reality, organizations will need to be very selective of how to deploy it: which tasks to assign to it, who should have access, and what guidelines, rules, and restrictions there need to be on usage.
The irony is that the pullback makes the allocation question temporarily moot. Intelligence can be technically achievable and commercially valuable while still being unavailable.
3. Task imagination: I became aware of the term “task imagination” through the AI Daily Brief podcast, which references a video by AI strategist Nate B. Jones. In his take on the Fable 5 release, Jones makes the simple observation that not many knowledge workers think about their work in terms of tasks that might take days to do. It requires a certain level of strategic thinking that may not actually apply to many roles. A model can work for days, but most people have not been trained to define a goal that deserves days of machine labor.
This suggests that those roles should either use less powerful models, or rethink their goals in ways that make the work more compatible with using a powerful agentic model. For example, an editor might call on the model to develop more granular editorial guidelines and style guides based on different article types (news, features, evergreen explainers, etc.). Reporters might build out investigative agents that don’t just surface data in document troves, but develop research plans based on leads and then execute on them by mining remote databases, filing FOIA requests, and other complex touchpoints that typically require human involvement.
It sounds good in theory, but realistically, many jobs have narrow definitions of what the work is, and there’s little motivation to go beyond that. Deploying the capabilities of a powerful model won’t necessarily help many of those jobs, at least as currently defined. That puts pressure on workers to imagine more ambitious tasks or risk being left behind.
Intelligence isn’t everything
Here is the paradox created by the pullback. A pause gives organizations time to build the governance, data practices, and strategic habits needed to use this level of intelligence responsibly. But task imagination is learned through use. Without access, people cannot discover which long-running assignments are worth the money, where agents fail, or how their own roles could expand around them. The pause buys time while taking away the main way to use that time well.
All of which is to say that stepping into a future where we’re working alongside agents has serious barriers beyond just capability (and political freak-outs over that capability). We restrict access to context so the tool (and its creators) doesn’t know too much. We limit how much we spend on models because we’re unsure of the return we’ll get. And many of us throttle our ambition with AI since our jobs simply don’t have a rich enough canvas for a model like Fable 5 to fill in.
And now there is a fourth restriction sitting above all three: the model itself may simply not be available.
For AI leaders, that could make ROI harder to prove. The strongest demonstrations depend on giving capable models real work, real context, and enough time to execute. When the best models are expensive, constrained, or suddenly removed, organizations fall back to lower-stakes uses that are easier to approve but less likely to change the economics of the work.
All of those concerns are valid, and don’t get solved by simply making better models. While advancements in security, infrastructure, and work redefinition will help us get past them, those are inherently slower than the rapid advancement of AI.
We have crossed one threshold and run into several walls. I suspect the story of Fable 5 will be looked back on not primarily as a step up in power, but the moment where the implications of that power pushed the limits of the systems meant to use it. Agentic models are certainly the future, but the present needs a minute to catch up.
The pause may be useful, but it comes at a price: experimentation is how organizations learn what this intelligence is actually for. For the moment, AI leaders may find that using frontier AI to its full capabilities is harder than proving those capabilities exist. We are out of the pure experimentation phase and into the reality-check phase, where access, cost, control, and utility matter as much as intelligence itself.
{“blockType”:”mv-promo-block”,”data”:{“imageDesktopUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/03\/media-copilot.png”,”imageMobileUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/03\/fe289316-bc4f-44ef-96bf-148b3d8578c1_1440x1440.png”,”eyebrow”:””,”headline”:”\u003Cstrong\u003ESubscribe to \u003Cem\u003EThe Media Copilot\u003C\/em\u003E\u003C\/strong\u003E”,”dek”:”Want more about how AI is changing media? Never miss an update from Pete Pachal by signing up for \u003Cem\u003EThe Media Copilot\u003C\/em\u003E. To learn more visit \u003Ca href=\u0022https:\/\/mediacopilot.substack.com\/\u0022\u003Emediacopilot.substack.com\u003C\/a\u003E”,”subhed”:””,”description”:””,”ctaText”:”SIGN UP”,”ctaUrl”:”https:\/\/mediacopilot.substack.com\/”,”theme”:{“bg”:”#f5f5f5″,”text”:”#000000″,”eyebrow”:”#9aa2aa”,”subhed”:”#ffffff”,”buttonBg”:”#000000″,”buttonHoverBg”:”#3b3f46″,”buttonText”:”#ffffff”},”imageDesktopId”:91453847,”imageMobileId”:91453848,”shareable”:false,”slug”:””,”wpCssClasses”:””}}
