Chicago's heat list changed who got arrested, not who got shot
The department scored everyone it had arrested since 2012 on a scale to 500, then handed the top scores to the young Black men its officers already stopped most.
One line in the system prompt turns off the bullet points
The bulleted, bold-headed answer is a post-training habit, built mostly by a stage that rewarded structure and length largely regardless of whether they improved the answer.
Both sides of the AI takeoff argument lean on the same measured trend
METR measured AI's reach on coding tasks doubling every seven months for six years, and the same line is the continuity camp's proof of a smooth climb and the fast camp's runway to a self-improving machine.
phi-1 matched code models 12x its size, but mostly on problems like its training set
Microsoft built the 1.3-billion-parameter model from about seven billion tokens of code that GPT-4 filtered and GPT-3.5 wrote, and it reached 50.6% on the HumanEval coding test.
Humanity's Last Exam only admits questions today's models fail
Its scores start near zero by design and climb fast, which is how a 26% run using browsing tools got read as the frontier of human knowledge.