spottedmarley

7 ms·

91.
by spottedmarley 2mo ago
The most disgusting humans on planet Earth (politicians) were appalled by the racist computer program. 'That is NOT our approved form of racism. This AI is evil, just like white people!'
92.
by spottedmarley 2mo ago
Elon has mentioned that future versions of Grok will write binary code directly. Once that happens, programming languages won't matter much to an AI.
93.
by spottedmarley 2mo ago
Nice. I actually have a custom command written: /speak-normal Can you guess what it does? Lol
94.
by spottedmarley 2mo ago
Micro-coking it's how code used to be written before AI
95.
by spottedmarley 2mo ago
Misaligned teenagers could too (way more scarier)
96.
by spottedmarley 2mo ago
' We need to make this big pile of turds more expensive. ' -MSMarketingTeam
97.
by spottedmarley 2mo ago
Most of my inference already happens locally, actually.
98.
by spottedmarley 2mo ago
Good for him, somebody ought to.
99.
by spottedmarley 2mo ago
Hey, what is the name of that un-contacted tribe of humans living way out in the rain forest? I'm going to contact them and tell them Zuck says the future is theirs too.
100.
by spottedmarley 2mo ago
Shortly after Sanders' urgent plea to hyper-scalers to pause their AI development, crickets could be heard chirping across the entire United States.
101.
by spottedmarley 2mo ago
Sloplicious!
102.
by spottedmarley 2mo ago
Regular updates. What a crazy idea
103.
by spottedmarley 2mo ago
I tried the same excuse after I fell asleep at the wheel and ran over an old lady and her dog. 'I swear I wasn't driving. The car just did that all by itself.'
104.
by spottedmarley 2mo ago
LLMs are not intelligent
105.
by spottedmarley 2mo ago
Research and analysis
106.
by spottedmarley 2mo ago
I think you are being very dismissive of all the fine, hard work that poor AI had to do to create that supersized helping of slop for you to enjoy. Maybe we need an AI-appreciation week you know, give the little bastards some love.
107.
by spottedmarley 2mo ago
It's a Dell Pro Max GB10. I have two that I can bridge together to run larger models across the 256GB pool but I only bridge them when I am experimenting with larger models and they will run fine but at lower tk/s with really only
108.
by spottedmarley 2mo ago
For learning it's all you need. Pick a good small model, I would recommend Qwen3.5 there are models ranging from 0.8b params (will run on a smart phone) all the way up to 397b params .. I run the 122b param model daily with 128GB vram
109.
by spottedmarley 2mo ago
and Religious Leaders Say the End Times Have Begun The real question is why anybody listens when salesmen tell them anything
110.
by spottedmarley 2mo ago
Ok, well what about blowing up datacenters?
111.
by spottedmarley 2mo ago
Makes me miss my old Packard Bell PB16 .. or maybe I just miss the 90's in general .. or being young.. sigh
112.
by spottedmarley 2mo ago
On Linux I use ShotCut for video
113.
by spottedmarley 2mo ago
But then how would I take the photo I need to feed my AI-powered fecal analysis agent?
114.
by spottedmarley 2mo ago
I've never stopped using Sublime Text. There's no reason. It does everything and also nothing (in all the right ways). Welcome back
115.
by spottedmarley 2mo ago
Claude is MacGyvermaxxing
116.
by spottedmarley 2mo ago
I've been wanting this for a while. I've developed my own routine for doing this and can't wait to stop doing it!
117.
by spottedmarley 2mo ago
The internet is the same as it's always been. Sending packets of data from here to there. Everything else is abstraction built on top of that. All of the cool stuff is still there it has just moved off the web into other sectors or evo
118.
by spottedmarley 2mo ago
I built my own arena for benchmarking local LLMs on a suite of small projects intended to grade the models on workflows I am likely to use them for. You can view my leaderboard: https://spottedmarley.com/arena
119.
by spottedmarley 2mo ago
Bring back funeral pyres
120.
by spottedmarley 2mo ago
Equates to: Our world is full of advertising and marketing

More ›