Replying to @JaneBColeman
congrats. you've just announced to the entire world that you dont know how adaptive testing works.
Pat Azi
1,947 posts
Pat Azi
@azi_pat
Harvard Law '12 → GauntletAI → Sr. PM, Westlaw AI | Bringing law into the AGI era
- Your kid's public school says they're "proficient." They're lying. Texas calls as low as 35% passing. Iowa claims 72% of kids are proficient in math, real number? 27%. That's a 45-point LIE. Your straight-A student is 2-3 years behind and you're paying for the cover-up.
- Replying to @samaI don’t care about free or plus users. What do you have for those of us paying $200/month
- Replying to @OpenAIIt's about time. Stop treating me like a second class citizen or I will take my $20/month elsewhere. Final warning.
- Replying to @sciencegirlSingle mom, negative account balance, starving kids, steals $18.75 worth of groceries from Walmart to keep her kids alive.
- Replying to @samaBetter drop something good soon. I pay grok 300/month, Gemini 125, Claude 100, and gpt only 20. I love AI and am very, very loose with money. The fact that you’re only getting 20/month from me says a lot. Step it up- now.
- Replying to @geoffreyhinton and @AGRobBontaWhen Hinton quit Google in 2023 to warn us about AGI risks, I thought 'If the godfather of deep learning is alarmed, we should be too.' But the same unwavering faith in neural nets behind his groundbreaking 1986 backpropagation paper also led him to wrongly declare in 2016 that
- Replying to @samaWhen can we expect GPT 6? I haven’t even started using gpt 5 yet and am already bored with it
- When Hinton quit Google in 2023 to warn us about AGI risks, I thought 'If the godfather of deep learning is alarmed, we should be too.' But the same unwavering faith in neural nets behind his groundbreaking 1986 backpropagation paper also led him to wrongly declare in 2016 that
- Replying to @VraserXYou’re just making things up. You have no idea what GPT 6 will be like.
- 2/3 Mark Chen: ‘Reasoning needs knowledge—you can’t RL blind.’ GPT 4.5’s the bedrock (token titan), priming GPT 5 (~mid-2025) for RL-augmented reasoning. Claude 3.7 flexes, but GPT 5’s going way bigger. This is AGI’s next leap—argue me.
