This is why I love the AI space at the moment. There is so much room for growth, exploration, and creativity often requiring little in way of financial investment. The space is like an ever evolving playground or sandbox constantly being built in by labs and individuals alike.
Cool stuff: This weekend I vibe-coded an entire multiplayer game built entirely around "prompting", think Minecraft but with prompts!
๐คฏ It's the first game I've made where my friends and I spent 5 hours straight playing, it was so fun and unique.
This map was 100% empty!
This is basically what I had hoped OpenAI's voice models would be like at some point as far as expressiveness is concerned (and maybe they'll still get their eventually).
Just started messing with ElevenLabs v4. WOW! SUPER impressed is my initial impression. Drunk voices? โ Can do ASMR (even sound effects inline)? โ Variety of effects (ie: not the same sounds for each tag)? โ I am sure it's not perfect, but close to my ideal "voice model".
I hear these hype posts about Dots (similar to new model hype posts) and I constantly wonder, are we using the same product? Is there like a Silicon Valley-version that gets updates faster or something lol. Then I remember how many of them get retweeted by OpenAI staff.
@joannejang I only ever think about her when people bring her(?) up, other than that I don't think much about her. Have heard things about them. Whether true or not, I don't know, but either way I don't really care enough to find out. The "viral" polls tell me enough.
@0xrwu I would mostly agree with this. Feels like: "If you have unlimited tokens what would you build?" "All the things!!!" It very much feels like designed products by the seat of your pants and as always I am reminded of the motto: "move fast and break things." Terrible idea.
Update: I can now see my Dot in Codex thanks to a Codex update (but at the same time that leads me to even further think this was rushed with very little extensive testing). Not a huge deal on its own, but not a good look (especially paired with the compute cuts to paid plans).
Why does my Dot appear to only have one chat thread in ChatGPT? Why cant I see my Dot in Codex (and am instead prompted to create a new Dot)? Technical issues aside, the more I use it the more half-baked OpenAI's Dots seems.
Sometimes higher cost or more intelligent models save you money in the long run as opposed to re-prompting dumber models to fix issues (potentially wasting tokens). GPT 6.1 Sol may be more efficient for many tasks but just some of the tasks I am trying to get done.
Hate to say it, but think I am going to re-sub to Claude. With Dots being in a bit rough shape, Astra 6.1 being supposedly cancelled, etc. Returning to Claude to try and get a little more work done. GPT 6.1 Sol is decent, but don't want to reprompt over and over to fix things.
@GavinNewsom Wow! Great job! Truly ground breaking work! Thank God the politicians are doing work that truly matters and not acting childish in any way.
Even if it couldn't initiate an actual phone call, I am not why it couldn't simulate one through the app. This is one of a few features that doesn't feel quite complete and more like "We really need a response to Grokbot and Muse for DevDay."
Seems Dots can't initiate phone calls... So I can only pretend to have a phone call with the agent (and must initiate)? Dots really seems a bit rushed in places in the implementation department. So far while I see the vision, I'm not sure it cooked for long enough.
Maybe to detect and recover when computer crashes happen on the local computer? I can think of a few others involving networking tasks, but those are really niche. I am not complaining about the feature though, just trying to understand the best use case for it.
One thing I don't quite understand yet is why I would ever use Dots to control my local computer. Anything I can think of is just an additional step with Dots; aka why wouldn't I just use Codex on the local computer? Hmmm... I am sure there is a use case, but seems niche. ๐งต
I sincerely hope Gemini 4 Argon is good. They've been cooking quite a while, so it would be kind of unfortunate if it wasn't. Only time will tell I suppose.
Say what you want about OpenAI or really any company, but bad mouthing a presenter or presentation because of an unforeseen technical issue is pretty fucked up. Complaining about the info presented? Ok, but the technical issues and presenter themselves? Not cool.
Some of the people complaining about a (temporary) live demo are the same people looking for any possible reason to criticize. Do I have criticism of OpenAI and their models? Sure. Am I going to write a harsh criticism of a presenter doing a live presentation or demo? Fuck no.
Polished and pre-scripted video may be another story or a tweet that a staff member had virtually limitless amounts of time to correct or polish, but a live moment on stage in front of a live audience? Hell no. I have been there. That shit is nerve-wracking when shit hits the fan