Nothing better symbolizes how humans are working at the mercy of machines than selecting squares with traffic lights.
Computers have been better than humans at this for many years; yet we keep bowing to their demands.
In case you were unaware: You’ve been training ’AI’, for free, all this time. 👇
Business Insider article ran this article in 2013 showing video of software that can bypass Captchas:
Tiny Startup Vicarious Is Creating A Human-like Brain That Runs On A Laptop.
Please ignore the bizarre claim in that headline, haha. 👇
www.businessinsider.com/vicarious-ge...
Tiny Startup Vicarious Is Creating A Human-like Brain That Runs On A Laptop
Humans, on the other hand, have struggled to solve Captchas from the very start, and those of us concerned about accessibility have been upset about their widespread adoption, signing petitions to end their use. My first post critiquing Captchas is from 2006. 👇
axbom.com/captcha-kill...
Captcha: Killing accessibility
Are Captchas are used to distinguish humans from bots? The first Captchas could only be solved by humans with above average visual acuity and cognitive ability to interpret purposely distorted characters. That excludes a fair number of humans. Or forces people into multiple, frustrating attempts. 👇
You also have to wonder, as Captchas evolved into image grids, how many people across the world have ever seen a US fire hydrant? 👇
In the end, enough humans were willing to spend time solving the Captchas that they just became too profitable to shut down. Why? Because they meant Google could make you work for them, for free. 👇
qz.com/2009829/clou...
Why AI developers love those annoying CAPTCHAs
This Techradar report from 2018 goes into some good depth about how Captchas were used to make humans transcribe books, and later train AI software. 👇
www.techradar.com/news/captcha...
Captcha if you can: how you’ve been training AI for years without realising it
A study in 2023 (Searles et. al, An Empirical Study & Evaluation of Modern CAPTCHAs) concluded that bots are superior to humans across the board (speed / accuracy) when it comes to solving different types of Captchas. My only surprise is that this particular study is a bit late to the game. 👇
Bots Are Better at Solving CAPTCHAs Than Humans, Research Shows
Is all this a problem? Well, Cloudflare made a rough estimate to highlight how it’s likely that about 500 years(!) per day are spent/wasted on solving Captchas. 👇
blog.cloudflare.com/introducing-...
Humanity wastes about 500 years per day on CAPTCHAs. It’s time to end this madness
Make no mistake, having difficulty solving Captchas has long been a human trait, while bots have been becoming more and more competent at solving them for a over a decade.
The opposite of what we are taught to understand what Captchas are, and what they are used for. 🚦
This thread pulled from parts of my newsletter in February of last year:
axbom.com/chatgpt-surp...