wsgeorge

In the immediate aftermath of GPT 3.5 and the LlaMA leak, Github Copilot, there was a lot of denial (negative sentiment, barely concealed rejection etc) about the world we were being ushered into.

Then there was a lot of movement by governments and pressure groups to legislate.

Post GPT-5, Grok4 et al, have we ridden the crest and gotten to the other side now?

Today OpenAI lets you try ChatGPT even before you logging in. Meta surfaced their models on WhatsApp. You can use a bit of Grok and Perplexity without giving up any personal information.

Huge contrast from Q1 of 2023, when LLamA leaked; when you had to go through hoops just to test drive an open model in a chat setting.

I'm just glad the market is more open now. :)

Short rant, but there's something awful about being blocked after authenticating by a puzzle - not one of those smart CAPTCHAs - that requires you to prove you're human just to log in to one of the most powerful general purpose AI services in existence today.

People don't complain enough about these tests, but I expect them to come to the fore as AI becomes more of a thing and users begin to feel like they're being treated as second-class netizens just because they're human.