Deedy slams WIRED's AI detector story: ad hominem over actual testing
deedydas · x · 2026-09-03
Deedy (deedydas) published a pointed critique of a WIRED article about an AI text detection product, arguing the reporting fails basic standards.
His main complaints:
- Instead of testing the product, citing research that actually evaluates it, or offering an informed opinion on AI writing, the author spent five paragraphs character-assassinating the founder, apparently feeling slighted;
- The article quoted a non-technical professor in a made-up field ("critical AI literacy") from an obscure university, based on a study that never used the product;
- It claimed bias against people of color and neurodiverse users without substantiating what bias means;
- It misquoted the product's accuracy after humanizer processing (it is not 4%).
In the original reply he quipped that the article was so badly written it might read better if AI had produced it. The thread is part of an ongoing industry dispute over tech media coverage of AI detection tools.
More from Fun
- US Open credentialed ~100 creators as media, then its umpire scolded them courtside — aakashgupta · 2026-09-03
- Overheard on a flight: 'helping my son dog sit — they're training a new model' — divy93t · 2026-09-03
- 'Twas the night before GPT-six-mas: community memes ahead of next OpenAI release — gpt2chatbot · 2026-09-03
- Old architecture debates keep resurfacing as 'discoveries' on X, researcher laments — A_K_Nain · 2026-09-03
- Dev jokes he wants a --dangerously-skip-permissions flag for his Grok bots — IndraVahan · 2026-09-03
- RSA-260 Factored, Setting a New World Record for General-Purpose Number Factorization — marvinvonhagen · 2026-09-03