
As a generative AI training model is exposed to more AI-generated data, it performs worse, producing more errors, leading to model collapse.

I am a technical and strategic founder developer experienced in web3, software engineering and building startups over the past 10 years.
I operate at the senior software engineer to CTO level, and am able to develop a company's "zero to one" product.
I’m ashamed to say I’m still using Reddit
Perhaps subscribe to communities that aren't for porn!
I think the problem with your suggestion is that NSFW-ness more of a spectrum.
Meanwhile both are SFW if you're in your pyjamas and working from home.
It’s finally happened, I’m devastated
This is so much more impactful than the Reddit blackouts. I feel a great loss.
I’m less driven by philosophy and more driven but whatever is available to use via my phone apps
Reddit quality has markedly deteriorated in my opinion.
Maybe it’s just my imagination.
Basically everything works but the content is less engaging!
Not for us Liftoff savages sadly
When in Rome…
I had a burst of activity on Lemmy and then slowly wound down to mostly lurking.
Not sure why. Maybe Apollo dying tomorrow will kick me again to participate
Some people love it, namely the ones that have most to gain.
So business owners, and extremely attractive waitstaff
It’s WAY faster than sharing an instance with normies. Think of it as your own personal caching server.
As for payment, you sign up to a VPS provider and give them your credit card, and pay for the usage.
Nice work! Yeah the database is king in these lands
Are you skilled with SQL? Everything should be working off the DB state so you just need to sift through and edit your user record
I'm talking about Loopring itself
This is a great point.
I've recovered from my Bitcoin maxi days, so I hope things move forward in the industry (namely, the normies learning about ETH)
Ah that would be make sense, but most people wouldn't see the point in running a node. People automatically think of "mining" or "validator"
I tried to like Loopring, but their L2s were hardcoded circuits rather than zkEVM which the Polygon and Matterlabs team (and Starkware to a lesser extent) are pushing ahead with this year. Allowing the community of third party developers to contribute value (sound familiar Reddit?) is going to make the whole L2 space to gangbusters.
As it stands now you can't do much with Loopring except what the first party devs have built, which is basically a standard excahnge.
DeFi has added some real value for both project creators and users, not just for myself. I'm not talking about the 'money making, profit driven' side of value either, but utility, capital efficiency, flexibility, new financial primitives as well.
All sorts of crazy innovation that you shouldn't just hand-wave away as a scam or redundant.
I guess it depends on your time-frame and your risk profile. BTC/ETH over the medium term has done much better than pretty much any other asset in the world. Altcoins didn't even factor into my brain when I made that comment.
Mastodon should just leave us alone!
A lot of controversial comments. Here are some of my observations:
I mean, most Lemmings (lol) hang out in Beehaw anyway, so centralised fediverse is already here
As a generative AI training model is exposed to more AI-generated data, it performs worse, producing more errors, leading to model collapse.
10 reasons MVC frameworks aren't dinosaurs but sharks
CUDA full GPU acceleration, KV cache in VRAM
This PR adds GPU acceleration for all remaining ggml tensors that didn't yet have it. Especially for long generations this makes a large difference because the KV cache is still CPU only on master ...
Go, Rust and the cost of brain power
2018-09-11 - Go, Rust and the cost of brain power - Thoughts and Ideas
An old blog post of mine from 2018 that did surprisingly well