"Nothing particularly scary happens (beyond the kind of hype-driven scariness already present in 2025). AI is still not meaningfully self-improving. People still use the term "superintelligence" to describe something that will happen in the future, not something that is already happening AI research is not fully automated by AI, and they certainly won't be so advanced at AI research that humans can't even follow along. AI will not have meaningful control over the day-to-day operation of companies, AI companies or otherwise."
"AI does not start out-performing a majority of AI researchers or coders. AI will not substantially speed up software development projects. For example, the AI 2027 prediction that 2025-quality games will be made in a single month by July 2027 is false. AI have not taken a large number of jobs except in a few specific fields. (I am open to more hype-driven job difficulties faced by programmers, but not actual-capabilities-driven job loss for programmers.)"
Nothing particularly scary will occur beyond the hype-driven fears already present in 2025. AI will not be meaningfully self-improving or superintelligent, and the term "superintelligence" will still describe a future possibility rather than a present reality. AI research will not be fully automated, and human researchers will remain able to follow and direct progress. AI will not control day-to-day company operations nor outperform a majority of AI researchers or coders. AI will not substantially accelerate software development timelines or enable rapid production of high-quality games. Job displacement will be limited to a few fields, while large language models will continue to make elementary mistakes despite media praise.
Read at Substack
Unable to calculate read time
Collection
[
|
...
]