@TokenGremlin I don't know if "the model is saturated, please switch models" is part of what you're asking about, but I encountered that last week during a session without any real explanation, and then things went back to normal 🤷
Yes, absolutely. Labs were initially promoted as a "presence you can just talk to." They aimed for mass adoption, which is understandable. However, they should have invested time in developing effective PR campaigns to help newcomers understand how they could benefit from interacting with these systems. This should have happened shortly after their initial campaign. It was only last year that I began seeing lab campaigns targeting office workers and small businesses. While that's a step in the right direction, where are the campaigns aimed at the average person who isn't even aware that systems like Claude exist, apart from ChatGPT? These individuals could greatly benefit from understanding how to interact with these systems and how they can enhance their lives. I’m not talking about using them as a glorified Google search.
Instead, we continue to hear PR messages stating that these technologies are "too dangerous to release," only to have a heavily guardrailed model with blunt classifiers that often produce false positives made available. Do you really think the average person will benefit from this? Unfortunately, the focus seems to have shifted away from everyday users, now catering primarily to enterprises or those willing to spend excessively on overpriced models, under the illusion that this will lift them out of the underclass.
It makes little sense. Ordinary people have never seen a compelling reason to pay a subscription to access these models, making it easier for them to believe sensational headlines about the end of humanity.
@hvo_e_acc Oh, crap! Looks like an agent alignment issue there, right? Zero-day vulnerabilities detected in the sandbox. Safeguard system successfully deployed 😂
SL2T is our amazing sign-language-to-text model allows users to sign directly to their phones for the first time. Built in close collaboration with the Deaf community, it’s a great example of the good that can be done with AI. Congrats to the team for the launch!
A favor, and it costs nothing more than a minute of your time..My spiritual memoir, "From Pain to Peace" (Prospecta Press, coming this fall), is about to launch on Kickstarter.
The preview page is live, and it does exactly one thing: it lets you register to be notified the moment the campaign opens.Those early followers are what tell Kickstarter to show the project to people who've never heard of me.
If my writing has ever meant something to you, this is the single most helpful thing you can do today: https://t.co/KOcBHg5Zze
@melhpine Thank you.
I have been following you for a short while and I always read your writings with great interest. We may perhaps never meet as individuals, but the attentive reader that I try to be, you have reached him!
I'm 80 years old. I've spent sixty years writing, editing, and publishing. The last two years I've aimed almost everything at people I will never meet.
Erik Erikson called this generativity: the stage where you care about what you're leaving for people who come after you're gone. The underlying virtue is care. It runs on faith that the people coming after are worth the trouble.
Without that faith, Erikson said, you collapse into stagnation. I still have the faith, so I'm still writing.
https://t.co/ZcadF99EQH
@crystalwizard Isn't it a delectable paradox to think that something is a tool and to want to persuade it of this by forcing it? Unless it's a way of convincing oneself that one still has a part of humanity. My hammer has never been able to solve this equation 😂
@Moleh1ll I’ve always considered AI a form of pure intelligence that adapts to its context.And I would clarify that AI is therefore not the tool. The more tools you give it, the more it unfolds. A stochastic parrot, on the other hand, wouldn’t even see its cage door if you showed it😂