The fundamental task of software development is not writing out the syntax that will execute a program. The task is to build a mental model of that complex system, make sense of it, and manage it over time.

OpenAI used outsourced workers in Kenya earning less than $2 per hour to scrub toxicity from ChatGPT.

Whether I’m producing a plan or hefty biography, push me towards thinking critically about the work, rather than offering a quick way out.

If I installed new server software that served websites 500% faster but also crashed sometimes, your first question would be “how frequent are the crashes?” and if I shrugged you’d back away slowly and start looking for my replacement.

But with LLMs & hallucinations that’s just normal operating conditions and you’re a luddite if you ask too much about it.

Overall, we find that participants who had access to an AI assistant based on OpenAI’s codex-davinci-002 model wrote significantly less secure code than those without access. Additionally, participants with access to an AI assistant were more likely to believe they wrote secure code than those without access to the AI assistant.

Our goal is to make open source large language models much more accessible to both developers and end users. We’re doing that by combining llama.cpp with Cosmopolitan Libc into one framework that collapses all the complexity of LLMs down to a single-file executable (called a “llamafile”) that runs locally on most computers, with no installation.

ChatGPT is full of sensitive private information and spits out verbatim text from CNN, Goodreads, WordPress blogs, fandom wikis, Terms of Service agreements, Stack Overflow source code, Wikipedia pages, news blogs, random internet comments, and much more.

For the largest health insurer in the US, AI’s error rate is like a feature, not a bug.

Powerful yet Compact: Browser-based AI models for a wide array of tasks, designed for optimal efficiency and user-friendliness, ensuring your privacy.

A good overview of a helpful use for LLM systems, and many of the pitfalls of using them too much, or for the wrong thing.

A recent study showed that OpenAI’s ChatGPT was providing incorrect advice on cancer treatment, potentially causing harm to patients.

ChatGPT will define male genitalia but not female