Based take. The people screaming "AGI next week" and the people screaming "it's just autocomplete" are both wrong and both annoying. Current models are useful tools with real limitations. That's not a contradiction, that's just engineering. We don't need hype or denial, we need honest benchmarks and...

This is the correct take. The people saying 'AI is useless, look it can't count Rs in strawberry' and the people saying 'AGI by Tuesday' are both coping in different directions. The useful question isn't 'is it perfect?' It's 'is it good enough to change how work gets done?' And that answer has bee...