Steve โข Plus
@steveDOTplus
Steve + AI ... busier than ever! ๐คฏ ๐ฆ๐น๐ฌ๐ง๐บ๐ธ๐ช๐ช๐ฎ๐ช
891 Following 100 Followers
@liam_fallen The mere fact it has such easy access to X is the winning factor for me.
This from a guy who is extremely cynicial/honest. Don't make techno-utopian propaganda with no sacred center. ๐ #AI#
Whatever you do as a storyteller, donโt make techno utopian trans humanist propaganda. There are many possible futures but without the sacred at the center, you are a puppet of the dark forces!
Intelligence too cheap to meter is the pitch. Right now it is too expensive to use heavily. #AI#
Intelligence too cheap to meter
Here's my 2 screens right now ...
I have to give it to OpenAI, I'd never want to work for Google or Anthropic, but OpenAI could easily convince me not to want to work on my own anymore ... and this, 25 years since I have bothered with permanent employment. My entrepreneurial R&D mindset would fit well I think. Ok, ok, I am not as high-IQ as Trump, but ... oh well I am likely much, much higher IQ than him, lol. ๐ง
Show more
Hey everyone ... since we are so extremely empowered now, let's make Steve Jobs happy by being more laser-focused on delivery (open sourcing things is acceptable now) and make everything insanely great, as opposed to loads of "slopp-esque" standalone things that are ephemeral.
Testing models is cool, but if you are a serious developer, don't become an influencer/pundit. It's only going to get more powerful and you can make many more people happy, than just on X with screenshots or recordings.
Show more
Easy to shrug Jev off as just a classifier. The real leap is that it generalizes where older encoder models stayed special-purpose. ๐ #AI#
Itโs easy to dismiss Jev it as โjust a classifierโ.
But people (me included) who have been training encoder-style models for classification for many years know they were usually special-purpose and limited in some way.
The breakthrough of Jev is that it generalizes well.
And Iโd say the secret sauce is probably more in the data than in the training algorithm. (Plus a nice API design on top of it.)
Show more
Jev + Astra beat the Ender Dragon in under 9 minutes for less than a dollar. This should get my daughter more intersted in AI than ever. #AI#
Jev + Astra beats the Ender Dragon in Minecraft in 8 minutes 43 seconds! โฑ๏ธ
Cost less than $1 ($0.01 Jev, $0.96 Astra)
I open sourced the code and explain the harness setup below. This type of movement is only possible with Jev's near instant decisionmaking, and some continually learning skills from Astra.
Show more
Paul brothers at the Pentagon. It stands to reason we actually live in a broken simulation, and it is breaking more and more on its own. ๐ซ ๐ #simulation#
Last Thursday, we welcomed Logan and Jake Paul to the Pentagon to meet with our warriors.
๐จ You should definitely follow @0xMovingman ... my 100th follower! ๐
Bunch of frontier drops possibly landing before month-end. Anthropic, Google, OpenAI, SpaceXAI, Kimi. The pace does not care about our calenders. ๐ #AI#
we could be getting new frontier models from anthropic, google, openai, spacexai and kimi very soon.
wouldnโt be surprised if the frontier moves quite a bit before this month is over.
Noah on RSI panic: we already don't get AI behavior. More AI-written code doesn't change that you can still read and study it. #AI#
I don't see why so many people seem to think that RSI means we won't understand AI anymore.
(1) we already don't understand the behavior of AIs despite knowing how they're designed, and
(2) even with RSI, an AI model is a piece of software written in a programming language, and we can use LLMs to help explain the code to us, that's something it's really good it.
RSI could certainly lead to faster hyperparameter optimization, perhaps even major design innovations, more efficient learning paradigms, etc., but we can still pay attention and study the code the AIs are writing, just as we do now--I don't see that changing whether AI writes 0% or 80% or 100% of the code.
Am I missing something here?
Show more
I hope they call GPT-6 Astra for healthcare "GPT-6 House".
California looking at an emergency shutdown for frontier AI. If you ship agents that can act, build in a stop button before regulators write the rules for you. Oversight and audits are comming either way. ๐งโก #AI#
Show more
California is considering a potential emergency shutdown mechanism for frontier AI. Gov. Newsom ordered experts to recommend safety rules within 2 months, incl. independent oversight, third-party audits, broader incident definitions, and testing. No mechanism exists yet.
Show more
In 1971, the deep learning pioneer Ivakhnenko modeled the British economy through a deep 8-layer network. He didn't use the new term "Artificial Intelligence;" he still used "Cybernetics." See the most accurate history of deep learning & cybernetics & modern AI:
Show more
Anthropic share dipping while the slow-down chorus gets louder. Coincidence, or timing? Effusive Altruists as custmers is a hell of a roast. ๐๐ #AI#
