AI Pulse · Frontier Labs Watch

Anthropic is losing goodwill fast among the developers and founders who used to champion Claude. Two things triggered the shift. First, the company launched Claude Design, a tool that competes head-on with Figma. Figma’s CEO went public, saying Anthropic wasn’t “consistently candid” about the move. Second, users report that Claude Fable 5 started giving worse answers to AI development questions, the exact use case many engineers relied on it for.

Now people are pulling workloads off Claude and moving to cheaper open-weight models from rivals. The complaints aren’t just about performance. Anthropic also skipped signing an industry letter backing open models that OpenAI, Google, and more than 70 other labs endorsed. That silence stands out when trust is already fraying.

This matters because developer loyalty is fragile. If your team picked Claude because it was the safe, thoughtful alternative to OpenAI, you’re now watching it compete with your design tools and degrade on technical queries. The cost difference with open models makes the decision easier. When a company stops feeling like a partner and starts feeling like a competitor or a black box, switching costs drop fast.

For anyone running operations that depend on a specific model, this is a reminder to keep your AI stack flexible. Locking into one provider’s API without a backup plan means you absorb their strategic pivots, their performance drift, and their relationship choices. An AI command centre that routes tasks across multiple models based on cost and quality lets you shift when one provider stumbles without rewriting your workflows.

The backlash is still building, according to the original report. Whether Anthropic can repair the trust or whether this marks a permanent shift toward open alternatives will show up in the next few months of usage data.