Go on, try out the Delhi Turing Test

In case you didn't know via neat dissemination and/or propaganda platforms, this week, Delhi's hosting the AI Impact Summit. It's the perfect setting for a Turing Test - not to check if machines can pass off as humans, but to check if humans can get away by pretending to know about AI. When you hear a panelist solemnly declare today, 'We must synergise quantum blockchain with generative intelligence,' you'll know he or she has read three buzzwords in this paper this morning. Real experts are muttering about gradient descent and hallucination rates, only to be drowned out by someone insisting that AI will soon replace 'all jobs, including politicians'. (Spoiler: AI still can't nail a punchline without it sounding like a dad joke.)

The summit promises 'impact', but the real impact will be on coffee machines, which must withstand the stress of fuelling 500 conversations that begin with, 'Actually, ChatGPT is just autocomplete'. The true test will be spotting the difference between those who can explain neural radiance fields and those who think it's a spiritual thing. So, let's run the Delhi Turing Test:

If you say, 'AI is the new data,' you're bullshitting.


If you say, 'AI models collapse under distribution shift,' you're probably legit.

If you say, 'AI will soon achieve consciousness,' you're a Keanu Reeves fan who should be at a comic con.