OpenAI offers up to $445,000 to manage recursive self improvement
OpenAI is gearing up to create AI systems that can actually make themselves smarter, and they are hiring for it, offering salaries up to $445,000.
The goal? To manage the risks of "recursive self-improvement," where AIs start training better versions of themselves with less human help.
This move comes as both OpenAI and Anthropic are making big leaps in AI tech.
OpenAI role addresses data poisoning threats
The new job will tackle tricky issues like "data poisoning" (when bad data messes with how AI learns) and will work on making AI decision-making more transparent.
OpenAI is also exploring ways their coding tools could support or even replace some engineering tasks, all while keeping cybersecurity risks in check.
Anthropic is on a similar path, showing just how fast the race for safe, self-improving AI is heating up.