Thariq Shihipar & Anthony Morris Announce They're Leaving Anthropic, Weeks Before Its IPO

Two engineers closely linked to Claude Code, Anthropic’s AI tool that can carry out coding tasks on its own, have announced that they are leaving the company. They made their announcements within one day of each other, just weeks before what could become one of the largest public stock market listings in history.

Thariq Shihipar, who used the name @trq212 on X and became one of the best-known people associated with Claude Code online, said he had left after “more than a year of working on Claude Code and seeing it scale larger than I ever imagined.” He added that he had “really loved working at Ant” (a common short name Anthropic employees use for their company). He also said he “deeply” believed in the company’s mission and, above all, in the idea of a “race to the top.”

Anthony Morris said his last day at Anthropic was the day before he announced his departure. “I created Claude Code on desktop because I believe everyone can be a builder,” he wrote. He added that every feature request and bug report from the community was “a gift that impacted millions of lives.” He thanked users and said, “the product wouldn’t be what it is without you.”

Who Are Thariq Shihipar and Anthony Morris

Shihipar joined Anthropic as a technical staff member on the Claude Code team. He became known for his work on the product’s skills system, ways of managing information and context, and what he called “dynamic” workflows, in which AI agents can adjust how they work. He often appeared on podcasts for software developers. In one appearance, he explained why he preferred writing technical plans in HTML rather than Markdown. In another, he described how the Claude Code team cut its system prompt, the instructions given to the AI, by about 80 percent to prevent fixed rules from limiting newer models.

Outside his engineering work, Shihipar was arguably the most recognisable person speaking for Claude Code on X, apart from Boris Cherny. He regularly responded directly to developers who reported bugs, requested features, or criticised the product. In practice, he became Anthropic’s public representative for Claude Code. This visibility also brought challenges. When Anthropic faced criticism over an incident involving hidden code included in a Claude Code release, Shihipar publicly responded to the backlash on the company’s behalf.

Morris, meanwhile, described himself as the person who created Claude Code’s desktop app. He built it around the idea that people who are not professional software engineers should also be able to write code. His posts suggest that he spent much of his time at Anthropic listening to users and responding to their feedback, with much of his work based directly on what the community requested.

Neither Shihipar nor Morris explained what they plan to do next.

A Pattern of Departures Ahead of the Listing

The departures come at an important time for Anthropic. The company confidentially filed paperwork for a U.S. initial public offering (IPO) in June. Reports since then have said it is aiming to go public as early as mid-November, with a possible valuation of $1.8 trillion to $2 trillion. If that happens, it would be the largest public offering ever, beating SpaceX’s record fundraising of $75 billion earlier this year.

Anthropic has historically had one of the lowest rates of employees choosing to leave among major AI research companies. Some commentators have partly linked this to the stricter limits Anthropic places on employees selling their company shares, compared with rivals such as OpenAI. Several senior OpenAI executives have left ahead of that company’s own planned public listing. Anthropic is also the only AI research company that still has its entire founding team in place. An upcoming IPO could change this situation significantly because employees whose shares have vested would finally have a clear way to sell them and receive money. This raises the possibility that Shihipar and Morris could be the first of many employees to leave.

The timing is also worth noting because Anthropic has been actively hiring new talent, including specialists from rival companies’ chip and hardware teams. At the same time, it is losing two developers who played major roles in one of its most important products.

Claude Code has become a key part of Anthropic’s growth story as the company prepares for its IPO. The company has said internally that it now uses Claude to write almost all of its own code. It has also presented Claude Code, which can carry out coding tasks with some independence, as an important part of its pitch to business customers and investors. Losing two engineers who were closely associated with building the product and representing it publicly, just as the IPO roadshow approaches, creates an image problem Anthropic would likely have preferred to avoid.