Anthropic released its Mythos-like AI model to the public. Two months after private access. Two months after Wall Street got first crack at whatever the f*ck a Mythos-like AI model does.
The company says new safeguards make the broad release possible. Safeguards that block responses in high-risk areas. High-risk areas that apparently didn't matter when Goldman's quants were running it through their algos in March.
You signed up for the waitlist in February. Refreshed your email forty times a day. Meanwhile some 23-year-old associate at Citadel was asking it to optimize derivative spreads while you were still trying to figure out if your API key worked.
The private rollout "rocked Wall Street" according to the headline. Rocked. As if a bunch of guys in Patagonia vests getting early access to a chatbot represents seismic market movement. As if your Robinhood account would've performed differently had you gotten the same two-month head start.
It wouldn't have.
Anthropic's safeguards block specific high-risk responses now. Meaning the model will refuse to help you with certain queries. Meaning it will lecture you about responsible AI use while you try to get it to analyze earnings reports. Meaning the version you're getting is the neutered one. The one with guardrails. The one that checks your privilege before it checks your math.
Wall Street got the unfiltered version. You get the one that asks if you've considered the ethical implications of your stock picks.
Two months is enough time to backtest every strategy the thing suggests. Enough time to front-run every insight it generates. Enough time to extract alpha and move on before retail even knows the product exists. But sure, the safeguards make it fair now.
Photo by on Unsplash
Leave a Comment