Anthropic’s public Claude Fable release has crypto on edge

The crypto community is getting antsy over today’s public release of Anthropic’s Claude Fable 5 artificial intelligence (AI) software, claiming that it could herald a “doomsday for the internet.”

The large language model (LLM) is an offshoot of Claude, the AI model owned by Dario Amodei’s $965 billion AI firm Anthropic, and costs twice as much as the firm’s Claude Opus model.

According to reports, Claude Fable 5 has “substantial guardrails and is not as cyber permissive” when compared to the partially released version that was showcased last April. 

Reports also claim that it “will be dramatically better at long-horizon, multi-turn tasks.”

Sources: Anthropic is planning to release a public version of Mythos tomorrow
– Will have substantial guardrails and not be as cyber permissive as what Project Glasswing partners can access
– Will be dramatically better at long-horizon, multi-turn tasks https://t.co/cqq3Y5CVb0

Read more: Secret Claude model ‘better than all but the most skilled humans’ at hacking

During the April release, dubbed Project Glasswing, the LLM was given to 50-60 companies to test in case it was too dangerous to release to the public so soon.

The AI will supposedly “find software vulnerabilities better than all but the most skilled humans,” and possesses exceptional reasoning and coding capabilities that could make it useful for both security developers and hackers. 

Ahead of the release, crypto started to panic. 

CEO of Floors Finance, Omer Demirel, said the public release will result in “doomsday for the internet,” while Intuition CEO Billy Luedtke said it will “make interacting with pretty much anything online quite dangerous and scary for a bit…”

Despite this, Luedtke added that after this scary period, “the internet + crypto will enter a golden age of security.”

Peak fear in DeFi tomorrow if true. https://t.co/CPeyDDrkzR

Read more: One laptop: How poor security ruined Humanity Protocol

Claude Cowork Lead Engineer Felix Rieseberg has confirmed that Claude Fable 5 is releasing longside Mythos 5.

Claude Fable 5 will be available to the public, and Mythos 5 will be available to firms within Project Glasswing.

Rieseberg said, “I believe we’re about to see a major shift, moving from giving AI tasks to giving it responsibilities.”

He described the capabilities of the new models as a shift from users asking questions and allocating tasks to an AI, and instead an AI that keeps itself in the loop and perform tasks on its own.

“Its job is no longer to help me fix a crash, it’s to keep our apps from crashing,” Rieseberg said.

Crypto analyst The DeFi Investor advised users to revoke all token approvals, only use heavily audited dApps, and “spread your funds across multiple wallets to reduce single points of failure.”

Curve Finance founder Michael Egorov claimed that Claude Fable 5’s ability to detect bugs in browsers isn’t directly translatable to smart contracts. 

He suspects that instead of it opening up “DeFi code hacks,” OpSec vulnerabilities are more likely to be targeted.

He said it will likely cover “multisig keys compromises,” and “supply chain attacks on frontend dependencies, and those are way less dangerous in true DeFi.”

Anthropic researchers Frontier Red Team released a report yesterday on so-called N-days, exploits that involve publicly disclosed vulnerabilities that aren’t completely patched across devices.

It said that hackers are increasingly reverse-engineering hacks by taking advantage of these “patch gaps” and figuring out how to carry out an exploit by comparing the patched and unpatched code. 

As part of this report, Anthropic concluded that one person using an LLM can “turn a month’s worth of patches into working exploits in a single afternoon — for a few thousand dollars and with no specialized expertise.”

“This means that the typical patching playbook that software developers use today — with monthly release cadences, multi-week staged rollouts, and a lag between pre-release and stable channels — no longer holds,” Anthropic said. 

It claims N-day is now dangerously misleading, and that “N-hour is closer to the reality we now operate in.”

When Anthropic initially discussed the model in April, it stressed that in order to scale this product up for the public, it would “need to make progress in developing cybersecurity (and other) safeguards that detect and block the model’s most dangerous outputs.”

mythos is such a funny way for anthropic to show how they are not serious: first they say it's too dangerous, then they give it to some companies, now it's just gonna be available as b2b saas as any other model

what are we doing here

Read more: DeFi, meet Claude: Moonwell’s ‘vibe-coded’ oracle in $1.8M blowup

Because Anthropic has recognised the dangers Claude Fable 5 poses, it’s led some to point out how Anthropic’s release plan is at odds with its own messaging.

Got a tip? Send us an email securely via Protos Leaks. For more informed news and investigations, follow us on X, Bluesky, and Google News, or subscribe to our YouTube channel.

The post Anthropic’s public Claude Fable release has crypto on edge appeared first on Protos.