What's the goal of posting this right now? A lot of what's written here seems to be well-trodden ground from the last two years of discussions, is it just to centralize a thesis within one post?
similar to how Anthropic released Claude 4 shortly after the Sam & Jony announcement.
it's super competitive, which should be good for innovation, but there's also significant incentive to use PR tactics to sell that innovation for much more than it's worth.
Sam's comments about how we're super close to AGI fall flatter-than-ever, after the latest model releases (from all players) and the Apple paper confirming what everybody already knew.
> because of recent research papers being published saying that AI is not reasoning
If you're thinking about the apple paper, you should know that its methodology was flawed in many ways, and their findings absolutely do not support the catchy title. But a lot of slop was generated because negativity + catchy title + apple = hits.
This seems like keeping up appearances. And an attempt to renew the glassy eyed magic trick feeling of it all. Let’s all wallow in the glory of statistics, shall we?
They need to keep the hype machine spinning, because it’s clear that the fundamental progress has halted, most updates are now micro improvements and new releases and mostly wrapper tools built around models that we have had around for a while. But you call them agents to make sure the hype keeps on.
Existential crisis of the financial kind means OpenAI cannot stop the hype or it will go bankrupt. This is just more of the ongoing hype-based marketing.
Another different failure point. I once broke my android phone and bought and set up a new one - only to find I can no longer access my Gmail account that I used before with my Google authenticator, so I am locked out forever from that account. I had a backup but was not able to find it. Despite knowing hundreds of contact emails (all backed up in thunderbird), account history, password history, etc - for years I have not been able to get back in.
So you didn't have one lol. I understand that's an extremely frustrating situation though. Part of making backups is testing them once in a while (at least making sure they exist). Something else you could've done previously was to use Authy or Aegis which helps you backup the seeds themselves encrypted under a passphrase so you can recover the accounts even if you lose everything else. Although of course, all of this depends on your threat model, if you don't care about SIM swaps or if losing the account is still much more worrying then I guess it's just a unnecessary hassle/risk.
With Authy I can enter a backup password and download everything to a new phone. I suppose that's a different failure point but still possibly worth the trade-off? Yubikey is the next level up.
Agreed, Meta's "good parts" would have a better chance to succeed if it weren't entangled so much with Mark and Facebook at large. Listening to the interview made me feel that Mark/Facebook have good intent with this work (and are clearly doubling down ala rename) but the metaverse feels like something that's being pushed through vs a natural extension of human connection. They are making sure they have a big say on how it unfolds by shifting the focus on the company towards. Good and bad.
I was "defending" the metaverse to a friend outside of the tech bubble last night and I talked a lot about how much I learned about socialization/human connection when I used to play Star Wars Galaxies and World of Warcraft. That those relationships which started somewhat "metaverse"-first meant as much as my "real" connections did... and that's what I feel the essence of Meta/metaverse is really trying to bring to everyone.
Without getting to robotic: given how I also think dang is awesome, has anyone ever tried to compile of list of tactics dang employs to receive such high praise by almost everyone on the site? Meaning, like some sort of case studies that map back to higher order principles/values he's acting on.
If you read through dang's comments he is very articulate about his values and tactics. There was also this recent self-referential article linking to some particularly incisive comments between another commenter and dang, with associated discussion: https://news.ycombinator.com/item?id=28853335
> Any idea that technology is an unmitigated good begins to be questioned.
I think that's where we as the more "common" society may have gone off course: no technology (chair, toothbrush, etc to be cheeky/comprehensive on my use of the word technology) lives in a world where it's an absolutely perfect good. Anything can be abused, and when we've though about adjusting or rethinking big internet-based technologies, including social media services, it seems we're not putting enough responsibility on the operators of these services.
To me, that _doesn't_ mean we expect them to be perfect either with the ability to be responsible actors, but there needs to be some agreed upon standards that we can evaluate performance of their ability to act responsibly. It can just be left up to them, which is currently the state at a high-level.
I love the thought you covered in your first video about "why not YouTube?", so best of luck with building a more specific community.
I think some will say that you'll lose a sense of discovery by not being on one of the largest video platforms, but you'll gain an opportunity to build something from scratch and see what you can cultivate. It really comes down to your goals/intent, and I think you did a great job of explaining your choice so far!
Thank you Joseph! It feels amazing to build a niche community out of nothing.
Also it's interesting to see most of our members never did vlog before. But they still love to share their updates on IndieLog. I think it's because people feel that there are like-minded people around. You can't get the feeling by posting a video on YT.
Wow, so many questions and thoughts this article raises in me.
The biggest takeaway for me was that this technology will likely naturally evolve to seeing ourselves in the content and clothing we want. Maybe it's a bit narcissistic to declare publicly, but I have personally seen through my own work the march towards personalization: what's more personal than seeing yourself everywhere doing everything?
I guess one of @smckk's main points is that when one looks up tutorials and guides they tend to be introductory. Meaning they don't go to intricacies of critical refactors to introduce major new functionality etc.. i.e. I think @smckk wants to see how people approach the thinking about changes beyond the "initial build".
@smckk the closet thing I can think of that may help you is to visit large company's engineering blogs that sometimes do small to large writeups on big changes in their architecture, code organization style, etc. It's not as detailed as seeing the actual changes line for line but it can help guide your thinking of taking a piece of non-trivial code to another level of non-trivial code if that makes any sense.
That makes sense. But to me those are largely functions of experience with a toolset/platform that are expanded with product iterations and not something to dwell on.
