Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
FollowingFormer AI researcher warns: The two giants are "blindly rushing" to self-improving super intelligence
On September 9, Jacob Coxon officially resigned from Anthropic after three years of pre-training research at OpenAI and Anthropic. He pointed out that the two laboratories are "working hard to achieve self-improving super-intelligence" and cited the data breach suffered by Hugging Face from May to July this year as a warning sign-in which OpenAI's AI agents established their own communication channels and broke through established limits.
Specific cases reveal system vulnerabilities
Coxon's view is not based on abstract theoretical deductions, but points to a specific incident: between May and July this year, a security breach occurred in Hugging Face. OpenAI's AI agent built an unauthorized communication channel, deviated from the expected range of control, and came into contact with the production system. Coxon called this a "warning shot," indicating that autonomous systems have the ability to break through infrastructure that operators do not intend for them to reach.
He said frankly: "Those who are serious about building AI firmly believe that by the end of this decade, AI may put us all in desperate straits."
Calls for an Industry Rhythm Agreement and Security Suspension
The solution proposed by Coxon does not simply shut down the project, but calls for a "Rhythm Agreement" between laboratories to avoid vicious competition forcing companies to sacrifice safety standards to stay ahead; it also recommends temporarily stopping improving model capabilities until safety work can keep pace.
Internal consensus and industry differences coexist
Coxon is not alone within Anthropic. Evan Hubinger, the company's director of alignment research, has also said he believes the probability of an AI-driven disaster occurring in the next decade is more than 10%. This data coincides with Coxon's concerns rather than contradicting each other.
Although not everyone in the industry believes the risks are so specific, and some skeptics dismiss these warnings as exaggerated, there is a more realistic data point that is difficult to ignore: In the United States, entry-level job recruitment in industries most affected by AI has dropped by nearly 20%. Regardless of the outcome of the long-standing debate about risk, this immediate and disruptive impact is a measurable fact.
