It's time to build the bay area megacity.
- rapid transportation (both intercity and intracity)
- high density
- high walkability
- enough housing to have 100 million+
- the home of 10+ trillion dollar companies
- underground hyperlogistics + robotic network
- capital of tech,
LFG!!! 🫡 finally finished devlog
#0
all actions & dialogue for every character is fully simulated, in real-time, with
@OpenAI
's GPT-4
and I condensed all the work so far into a 2-minute explainer, just for you 😉
YOOOOOOOOO! The conversation system is in & it's looking amazing!!
Completely generated in real-time, nothing pre-written, and 100% based on the character's personality, memories, and environment.
imagine getting access to the CTO of OpenAI's Twitter account, and wasting it on an obvious crypto spam instead of announcing GPT-5 and moving markets
🤡
It's time to build the bay area megacity.
- rapid transportation (both intercity and intracity)
- high density
- high walkability
- enough housing to have 100 million+
- the home of 10+ trillion dollar companies
- underground hyperlogistics + robotic network
- capital of tech,
@PiquantParvenu
I can’t hide the safari URL bar and it takes up a huge chunk of the screen, also swipe gestures are more annoying, and there are just a bunch of APIs Apple doesn’t give u access to like vibration
i don't know why, total moment of clarity right now...
Silicon Valley is such a meme, in a good way. I scroll through my X timeline, like posts by cat with glasses, anime girls, and ripped Jeff Bezos.
I have friends who are 20 raising millions to build startups, and they're
I made an open-source dashboard for monitoring
@OpenAI
's API. (e.g. response time, tokens per second, time to first byte)
Built w/
@nextjs
13,
@vercel
postgres, and
@Cloudflare
workers.
Pings each endpoint every 5 minutes and shows the result.
new startup founder strategy:
- focus on marketing & distribution day one on crazy impossible thing that can't be done yet
- wait for OpenAI to release the model that makes it possible
- profit?
my GPT agents are starting to look ... actually kinda good?
added:
- multi-agent simulation
- new character designs for variety
- auto camera tracking (could use some work)
- caching layer to help me not burn endless cash
i present to you: the ability to experience any story you want with a godly narrator and images to match
come try it out and burn my API credits!
(audio on)
It's an alert. Whenever the latency of OpenAI's API goes above a certain value, it's no longer efficient to test. When it comes back down it is, so I need to know when it breaks that threshold so I can remotely toggle my GPT agents.
i am far too addicted to X to be able to do the big polished release post, but as they say "if you are not embarrassed by the first version of your product, you've launched too late."
come make cool stories and pretty pictures on the rewrite of my AI storyteller, Bonfire.
LFG!!! 🫡 finally finished devlog
#0
all actions & dialogue for every character is fully simulated, in real-time, with
@OpenAI
's GPT-4
and I condensed all the work so far into a 2-minute explainer, just for you 😉
@StarveTheBeast3
@petrillic
i feel like I'm missing something
Netflix is on AWS and video is a few orders of magnitude more bandwidth than tweets.
Twitter has ~330M MAU and Netflix ~80M.
starting to realize a ton of industries are completely controlled by a cartel-like association or organization at the top that just stifles anything they don't like
genuinely speechless right now
told chatgpt to be a professor in a topic, and I asked it all my questions back and forth for an hour+ and I feel like I've learned soo much
The age of personalized tutors is going to be absolutely insane..
The OpenAI team is easily one of the most cracked teams out there, and if we see Sam, Greg & co come back with a fixed governance structure I think we’ll see the company accelerate even faster than before.
I truly believe OpenAI team may become stronger than ever in the next few months. The greatest risk was expansion and being spread thin, a shifting culture. Can’t imagine anything that would bring us closer together, nothing that would ignite more passion to continue the mission.
Introducing the DJI AGRAS T50. It has a 40 kg spraying and 50 kg spreading capacity, a spraying max flow rate of 24 L/min, and is ideal for heavy-duty agricultural operations that require precision and accuracy.
Learn more 👉
this is the coolest model in the world
- text, image & video can be used as input (all at the same time if you want)
- can create video at any aspect ratio between 16:9 and 9:16
- it can SIMULATE & play minecraft?!!!
holy shit. just wow
meeting people who will fly halfway across the world, sleep in a pod, and push past the impossible to chase their dreams is the definition of the American dream man
i love sf
ranting into the void here
solo-founding sucks, having no one to balance out the mood swings of the startup journey is damn hard
and finding the people that can fill that role is also hard.
@frantzfries
not sure how people expect Reddit to make money when
- their ads aren't served to anyone since everyone uses third-party apps
- their data is scraped for AI training left and right, because of cheap API access
no other platform has this many third-party frontends. I think
@MelindaBChu1
@planefag
@obversers
@andrewmccalip
You obviously don't know Andrew or Varda.
Varda has all the right tools to replicate this properly according to the paper, and the goal isn't to discredit their work but to see if it can be reproduced. If it can, then these folks are basically guaranteed a nobel.
Itd be
imagine games where you could act as a general, giving orders that propagate out to each platoon, or whispering rumors in a royal court
when you can influence NPCs actions, a lot of possibilities open up 👀
@ylecun
"Generation is very different from causal prediction from a world model."
I have trouble understanding this. With enough training data and computing, if a text-to-video model can produce realistic lighting, water simulation, and more, is that not a form of prediction?
ChatGPT can now see, hear, and speak. Rolling out over next two weeks, Plus users will be able to have voice conversations with ChatGPT (iOS & Android) and to include images in conversations (all platforms).
Over the last three months we have worked tirelessly toward the creation of Poseidon V1 - today it flew for the first time. Shortly we’ll officially announce what we’ve been up to, and where we plan on going.
@DavidSacks
Appeasement doesn't work. First, it was Crimea, now the eastern territories of Ukraine. If you concede them now, Russia rebuilds and goes further in the future.