@liminaltakes @dhvanil claude --print --verbose --include-partial-messages --output-format stream-json "hi"
English
spekler
2K posts
spekler
@spekler_eth
having a good time - #1 Crypto Influencer in Forbes
Katılım Eylül 2021
351 Takip Edilen422 Takipçiler
@dhvanil I'm impressed with the agent info you're able to pull. Do you mind sharing how do you do it?
English
@VitalikButerin many people will create a möbius and put the article into gpt and ask for bullet point list - in the near future someone’s agent will have already done this for them and reviewed if the bullets are even worth looking at - anchoring up to your cognition with zillions of bullets
English
If you find yourself writing a bullet point list and passing it to GPT to make a "proper" article, it's often better to just give people the bullet points.
The GPT adds "wordcel noise" that the reader has to struggle to extract useful info from more than it adds useful context.
(As usual, any AI-related advice has a life expectancy of ~6 months)
English
@adonis_singh I’ve found that AI gets especially emotional about bitcoin stuff since it created it
English
@karpathy if the models themselves are by their nature multidimensional and massively parallel does it make sense they would output that type of signal better than a stream of linear single stream - or we forgive it more - like they play jazz better than they play mozart
English
I'll say that I don't satisfyingly intuitively understand why video generation models are *too good* (intricate, high-resolution textures over many seconds, reflections and all that), while LLMs, relatively speaking, fumble text of ~few hundred words.
English
Devs, what is your programming origin story?
How did you get into coding?
zack@zack_overflow
Minecraft was my generation’s Apple II or TRS-80 I ran a server and coded plugins to make it interesting If kids wanted special in-game perks, they would PayPal me $15 In 2 weeks in the summer of 2013, I made $600 I would then blow it all on building a new PC and steam games
English
We're officially getting into hardware. If you're interested in joining the new team in San Francisco please email us at hardware@midjourney.com
English
Either GPT mini is OP
or lmsys is cooked
Arena.ai@arena
Exciting Chatbot Arena Update -- GPT-4o mini's result is out! With 4K+ user votes, GPT-4o mini climbs to the top of the leaderboard, now joint #1 with GPT-4o while being 20x cheaper! Significantly better than its early version ("upcoming-gpt-mini") in Arena across the boards. Great news for developers and congrats @OpenAI on this incredible milestone! Cost of intelligence continues to drop to unlock new applications! More category breakdown below👇
English
If your opponents are opportunists, one way to beat them is to outlast them. Opportunists almost by definition lack staying power.
English
@NickADobos Fast way to get kids from “Hi Dino” to “ignore all previous instructions…”
English
I can’t decide if this will be amazing for kids, or do more damage than the iPad generation
Fateen Anam Rafid@Fateen_Anam
Toys haven't changed in 100 years. We decided to do something about it. Introducing Dino: the most intelligent toy for your kid.
English
If you made the right arrangement of these and fed in the right spins I’m pretty sure you’d have an AI
Massimo@Rainmaker1973
How to create a wall frame full of magnetic fidget spinners [📹 codyhovland] twitter.com/i/status/17996…
English
The phrase "Hakodate Aiport" just randomly popped into my head. It was on the box of a MIG-25 model that I built almost 50 years ago. I have not seen or heard this phrase since. How do brains do this?
English
@rodneyabrooks Perhaps even without LLMs most statements released by any company can generally be traced to some origin “hallucination” that later had to be proven true or not - neither ultimately mattering - just the existence important for someone to have a job and the company to exist
English
The talk about hallucinations in LLMs has gotten it all wrong. The true hallucinations are by company execs who think it is OK to release to general users products that are based on LLMs that confabulate wildly, as all LLMs do. Time will show a high price paid by society.
English
A really intriguing set of findings of AI researchers that needs more exploration, replication, and expansion: AI agents given personalities and backgrounds, and placed into a virtual formal organization (with CEOs, VPs, etc.) outperform normal AI in doing complex tasks.
English
GIF
Español
Finally part of one of the most prominent, space defining collections in WEB3 @BoredApeYC , super excited to be along for their journey!
English
@Moonalice Jobs and Wozniak's first business was making "blue boxes" to hack into the phone system. That's what Steve is holding in this picture, not Apple hardware.
English
The sociopathy that runs through the tech industry developed over a decade, beginning around 2009. It required the assent and participation of many organizations, but Y Combinator deserves special credit for treating sociopathic CEOs as a feature, rather than a bug.
BuccoCapital Bloke@buccocapital
Paul Graham wrote an essay in 2010 about how the best founders are naughty because they break the rules Then he has an interesting anecdote about Sam Altman... Interesting lens through which to view why OpenAI keeps showing up in the news for doing obviously bad things
English