
babylon is a vampire π§ open source ai things & stuff. dev: https://teapawt.net/t.co/t6Rk2LlhpR https://teapawt.net/t.co/FQckz5sF4Q contrib: @NousResearch
the abyss
Joined January 2009
- Tweets10.3K
- Following83
- Followers628
- Likes6.2K
Tech site @builtin just published a piece about open-weight AI's "hidden risks." It read like a puff piece for the big labs; builtin.com/articles/open-weβ¦
1
1
18
If that's the threat model, the real risk isn't to you - it's to the people selling access for princely sums.
prasta.substack.com/p/spoileβ¦
1
9
For the record: if 9.4M shares were my fund's largest position - 19.2% of the book or $1.8B per the 13F - I'd be concerned too. But I'd put the number in the letter to avoid the appearance that I was doing this for reasons other than self-serving ones.
sec.gov/Archives/edgar/data/β¦
150
pruster retweeted
I am one of the people behind the 6,500 uncensored models the press now wants off Hugging Face.
Transformer News counted over 6,500 abliterated, uncensored and heretic models on Hugging Face in July. Their story is that these models could help cyber attacks. Hugging Face removed 1 model marketed for offensive cyber, and the reporter asks why the rest are still up.
A few of those are mine. I upload abliterated Qwen and GLM quants. My Qwen3.8 Flash-Next one got about 358,000 downloads in the last 30 days.
Abliteration removes the refusal direction. That is all. The model does not learn hacking from it. It stops lecturing you when you ask about a chemistry exam, a pentest you are paid for, or a dark scene in a novel.
The base models are already on millions of drives. Removing my repo removes nothing.
The real problem is somewhere else. Hugging Face is where almost everyone gets weights. Nvidia agreed to buy it, Nvidia just lost the CEO who promised to keep it open, and now the press asks why uncensored models are still there.
Download what you need. TODAY!!!
In other bad news for Nvidia: @stokel found that Hugging Face, which Nvidia recently agreed to acquire, appears to be hosting a load of models that could be used for cyber attacks.
transformernews.ai/p/huggingβ¦
Readers added context they thought people might want to know
This is a disingenuous take designed to be alarmist. All models capable of defending against cyber attacks are, by their very nature, capable of performing them. Hugging Face, itself, had to use such a model when OpenAI attacked them and then frontier models refused to help. fortune.com/2026/07/20/hugβ¦ finance.yahoo.com/video/hugging-β¦
Which open source projects had a fast rise and an even faster fall?
Not the slow fades. The ones teams adopted enthusiastically, built on, then quietly ripped out a year (or less) later.
I'm collecting historical examples. What's your pick? #opensource
14
pruster retweeted
This is a Zebrafish
It has the same amount of Synapses as Qwen 3.8 27B but lacks the ability to one-shot Minecraft in Three.js
Disappointing
no no no no no no makerworld.com/en/models/213β¦
19
pruster retweeted
Sometimes I think my wife forgets that I jumped a 10' wide creek on my BMX bike in 5th grade and landed it.
She would lock up that attitude if she remembered what I'm capable of.
pruster retweeted
You can now join the waitlist for the Hermes Agent mobile app to get in as we get more seats!
Thank you to everyone who traveled from across the world to NousCon 2026. We loved spending time with the people who make the Hermes community great.
As you all know, we released the Hermes Mobile beta at the event, and as anticipated it filled up quickly. We are working tirelessly to free up more room in the beta, and to get the app production ready for wider release. The feedback has been incredible so far, and we are excited to share it with the world soon.
In the meantime, please join the waitlist at hermes-agent.nousresearch.coβ¦














