[2025-02-12]

It’s reasonably hard nowadays to not follow AI developments, exciting as they’re - but certainly overwhelming. Math olympiads, Benchmarks, Agents - no threshold is safe.

Irony is - that AI companies are not at all safe as well and first movers do get pressured, it even seems to happen quicker than ever before.

At least as a discussion topic on tech forums which I look into with updated n-gram dataset.

First of all, referencing amazing blog post from 2021 by Abraham Thomas on the meaning of shortened timeframes on the market (esp. for VC-backed companies). It just tends to overshoot on exposure where even tiny slump in growth / some event / etc. can cause big outflows.

Case-and-point shared around on social, some AI products (wrappers too, oh the irony v2) are among the fastest by ARR growth ever.

While it’s a ‘Normal/Good’ metric for the case, what interesting is how actually little of a moat there’s.

Cursor itself seems to have defeated or being close to beat the OG of the new era AI products - Copilot.

Same goes for first chat - ChatGPT, not beaten yet and cadence is admittedly very strong, but among HN audience discussions it’s clearly not growing that forward.

Especially, of how fast the new entrants (like Deepseek) managed to get attention with a high-quality products.