AGI questions, answered plainly

There is no agreed answer, largely because there is no agreed definition or test. Some people argue current systems show early general abilities; others argue they fall short on reliability, reasoning, and learning new tasks. Any claim that AGI "has arrived" should say which definition it uses and how that was measured. See What people mean by AGI.

No. AGI usually means roughly human-level ability across many tasks. Superintelligence refers to systems far beyond the best humans in nearly every domain. They are often discussed together but are different ideas.

Nobody has the authority to decide. Companies, researchers, and commentators each use their own definitions, and some companies have commercial or contractual reasons to declare or deny it. That is why evidence and definitions matter more than the label.

Not on their own. A benchmark measures performance on its specific tasks. Scores can be inflated by test data leaking into training, and a model can do well on a test while failing on similar real-world tasks. Read How to read an AI benchmark claim.

"Bling" means flashy, showy jewellery. The site is about separating the shine of AGI announcements from the substance underneath.

No. It is against claims that cannot be checked. Real progress deserves clear evidence, and the hype check is meant to be applied to optimistic and pessimistic claims alike.

No. It measures how well a claim is supported with checkable evidence. A well-evidenced claim can still be wrong, and a poorly presented one can be right. It is a reading aid, not a verdict on the technology.