Hollinden

Anthropic launched Claude Fable 5 earlier this month as a publicly available version of its more advanced Mythos model family.

On June 12th, Anthropic said it received a U.S. government export control directive requiring the company to suspend access to Fable 5 and Mythos 5 for foreign nationals, including foreign-national employees inside the United States.

Anthropic then disabled both models for all customers to ensure compliance.

According to Anthropic, the government did not provide specific details about the national security concern in its written directive. The company said its understanding is that officials were concerned about a potential jailbreak method for Fable 5.

Anthropic has disputed the severity of the issue, saying the technique appeared narrow, involved a small number of previously known vulnerabilities, and did not make the model meaningfully more dangerous than other publicly available models.

That disagreement is now part of a larger story: how governments, AI companies, security researchers, and businesses will define responsible access to frontier AI.

What do you think: Did the government make the right call by acting early on a potential AI security risk, or should companies and customers expect more transparency before access to major models is restricted?

The accounting firm M&A conversation tends to produce two camps: firms that are actively pursuing a transaction and firms that are waiting to see how the market develops. Neither orientation is independently a growth strategy.

The firms building durable independent practices are making explicit decisions about service mix, client concentration, partner succession, and market positioning that strengthen the firm regardless of what the transaction environment looks like in any given year. The discipline required is indistinguishable from the discipline required to prepare for a transaction, because the underlying work is the same.

Independence by design means the firm's value is not contingent on external validation. It means options remain open.

What decisions is your firm making right now that would hold up under scrutiny in either direction?

https://hubs.ly/Q04jb0dS0

The restructuring of the accounting firm competitive landscape since 2023 has not reversed. PE-backed platforms have access to capital, technology infrastructure, and talent acquisition capacity that most independent firms cannot match directly.

What they cannot replicate is what well-run independent firms have built over years: senior partner continuity, client relationships with genuine institutional depth, and organizational agility that does not require a board approval cycle.

The independent firms that are competing effectively are not attempting to mirror the PE model. They are articulating what they offer that the PE model structurally cannot - and positioning that distinction clearly enough that the right clients recognize it before a conversation begins.

We wrote about what that looks like in practice.

https://hubs.ly/Q04jb4cz0

Firms that serve a wide range of clients across multiple industries and service lines have made a strategic decision, whether they recognize it as one or not.

That decision has predictable consequences. Clients in high-complexity verticals engage specialists because they require advisors who understand the specific dynamics of their industry - not competent generalists who will develop that understanding over the course of the engagement. Premium fees reflect that distinction. So does client retention, referral quality, and the speed with which a firm becomes the obvious choice in a pursuit rather than one of several reasonable options.

The generalist discount is real and it compounds. The path away from it is not a rebrand. It is a series of deliberate decisions about where to build depth, made early enough to matter.

Where is your firm already perceived as a specialist, even if that perception has not been formalized into a market position?

There is a meaningful difference between a firm that does concentrated work in an industry and a firm that has built a defensible position in it.

A genuine specialty practice has documented systems designed around the niche, staff trained in it with bench depth that is not dependent on a single partner, a referral network mapped to the specific vertical, and a market presence that makes the expertise visible before a prospect initiates contact. Firms that have claimed a specialty without building that infrastructure tend to discover the distinction quickly when a sophisticated client or referral source examines it closely.

According to https://hubs.ly/Q04jrrXz0's State of the Profession data, firms with clearly defined target markets grow 23 percent faster than generalists. That differential compounds. The practical question worth putting to your leadership team: where is your firm already billing above its average rate, and what does that reveal about where the real specialty lives?

We wrote about what it takes to formalize that position and build the platform around it.

https://hubs.ly/Q04jrplM0"

Prospective clients are asking AI assistants the same questions they used to ask search engines.

""What is a quality of earnings report?"" ""What should I look for in an advisory firm?"" ""Who handles M&A due diligence for mid-market deals in my city?""

The firms that show up in those answers - whether through organic authority or paid presence - are the ones clients find first.

Answer Engine Optimization (AEO) is the practice of structuring your content so AI systems cite and surface it. It is real, it is measurable, and it is one of the clearest competitive separators we are seeing right now between firms that have invested in it and firms that have not.

The question worth asking this week: when a prospective client asks an AI about firms like yours, does your firm come up?

One of the most direct pieces of advice we have heard from AI researchers at the university level: "Firms that aren't practicing BI shouldn't be focused on AI."

AI does not create insight out of thin air. It surfaces what already exists in your data. If your client records are inconsistent, your pipeline is managed in someone's head, or your CRM has not been meaningfully updated in two years - AI will not save you. It will amplify the gap.

The firms building competitive advantage with AI right now invested in clean, organized, and consistently maintained data before they started experimenting with the tools.

The infrastructure conversation is not glamorous. It is foundational.

Is your firm's data ready for what AI can do with it?

Read more here: https://hubs.ly/Q04hmmmF0

Today, we pause to honor and remember the brave men and women who gave their lives in service to our country. Their sacrifice, courage, and dedication will never be forgotten.

This Memorial Day, we reflect with gratitude on the freedoms we enjoy because of those who served and made the ultimate sacrifice.

"Do you use AI?" used to be a curiosity question. It is becoming a client evaluation question.

The firms that handle it well are not the ones with the most sophisticated tools. They are the ones who have thought through what they use, where it adds value, and how to say that clearly without overpromising.

Mid-market and regional firms have genuine advantages that large platforms cannot replicate: senior attention, long-standing relationships, and local context that a global platform cannot systematize.

Those advantages deserve to be named and positioned deliberately - not left implicit and unspoken.

What does your firm say when the technology question comes up?

On April 7, EY announced the global rollout of enterprise-scale agentic AI across its assurance practice. 130,000 professionals. 160,000 audit engagements. 150 countries.

The question we keep hearing from mid-market firms is: What does this mean for us?

The honest answer is that it depends entirely on how your firm frames the moment. Dismissal is a mistake. Panic is also a mistake.

We wrote about where the strategic opportunity actually sits - and why this is a positioning problem more than a technology problem.

https://hubs.ly/Q04hmDx-0