When ChatGPT 5 was launched earlier this yr everybody anticipated an enormous step ahead in its capabilities. Because it turned out, there was trivial enchancment in what it may possibly accomplish.
There’s a easy cause for this slowdown in progress: we’re operating out of knowledge, prompting a scramble, and billions of kilos of funding, to create artificial knowledge.
Nevertheless, there are two issues with tackling the information scarcity with synthetic knowledge.
Firstly, artificial knowledge has been proven to introduce errors. It may well perpetuate biases, mislead selections and result in leaks of delicate knowledge. Secondly, artificial knowledge essentially can not create any new intelligence.
When ‘new knowledge’ is created from outdated, it’s merely an echo chamber. Gartner has estimated that by 2030, artificial knowledge will probably be extra prevalent than actual knowledge, which means that we’re hitting some extent the place the event of AI has stagnated….
