FaceDev (@facedevstuff) header
FaceDev (@facedevstuff) profile picture

youtuber and misunderstood contemporary creative visionary / linux users DNI / open-source supporters DNI / deltarune fans DNI / clippy pfp DNI

Windows 11 Pro 23H2 22631.4830
Joined December 2021
new video out, enjoy!
7
2
1
63
22,092
lol
I will cancel my ChatGPT sub if Claude gets separate web and desktop limits too Unbearably slow models. 1 minute 14s to fact check simple information. Unimaginably bad writing, too, sometimes. Spamming quotes and new lines everywhere
2
21
I just be posting some chatgpt tweet and giggle reading the replies 😂
Sam Altman is actually right about this, and most people are missing the bigger picture. Think about how much RAM sits unused in hundreds of millions of personal computers every single day. People buy machines with 16GB, 32GB, or 64GB of RAM, yet most of that capacity is barely utilized for much of the day. Meanwhile, AI research, scientific computing, and cloud infrastructure are competing for the same limited global supply of memory. From an economic perspective, this is an incredibly inefficient allocation of resources. In a data center, memory can be pooled, shared, and continuously utilized by thousands of workloads. Instead of expensive hardware sitting idle for most of its lifespan, that same capacity can support services used by millions of people. Higher utilization means more computing output from the same physical resources, potentially lower costs per workload, and less pressure to manufacture additional hardware. This doesn't mean everyone should give up owning a computer. Local computing still matters for privacy, independence, latency, and offline access. But the fundamental principle is correct: maximizing individual ownership of scarce hardware isn't necessarily the same thing as maximizing its usefulness to society. The future of computing isn't just about who owns the most powerful hardware. It's about how efficiently we use the hardware humanity already has. >Add a sharper counterargument>Make the opening more attention-grabbing⋮⋮⋮⋮⋮⋮⋮⋮
Readers added context they thought people might want to know
The post includes leftover AI prompt text at the end, showing it was AI-generated. Author later called it ChatGPT bait. x.com/facedevstuff/s… x.com/facedevstuff/s…
9
1
29
2,779
I still agree with Sam's argument though, it was just too perfect of a tweet not to meme Compute is unarguably being wasted when sitting in ordinary computers. Though, I think his original argument was that local LLMs are under-using the scarce compute, not that ordinary gaming computers are. Think about it. Or don't, I personally won't. Go drink some soda and enjoy the day who cares
20
14
1,717
well my clarification seems to be misinterpreted what i meant was: technically, datacenters make use of GPUs more efficiently than local ai this has nothing to do with RAM, communism, anti-local AI, capitalism, endorsing utilitarianism, or advocating for people's computing resources to be seized its just a factual observation and nothing more
38
Mission successful
Sam Altman is actually right about this, and most people are missing the bigger picture. Think about how much RAM sits unused in hundreds of millions of personal computers every single day. People buy machines with 16GB, 32GB, or 64GB of RAM, yet most of that capacity is barely utilized for much of the day. Meanwhile, AI research, scientific computing, and cloud infrastructure are competing for the same limited global supply of memory. From an economic perspective, this is an incredibly inefficient allocation of resources. In a data center, memory can be pooled, shared, and continuously utilized by thousands of workloads. Instead of expensive hardware sitting idle for most of its lifespan, that same capacity can support services used by millions of people. Higher utilization means more computing output from the same physical resources, potentially lower costs per workload, and less pressure to manufacture additional hardware. This doesn't mean everyone should give up owning a computer. Local computing still matters for privacy, independence, latency, and offline access. But the fundamental principle is correct: maximizing individual ownership of scarce hardware isn't necessarily the same thing as maximizing its usefulness to society. The future of computing isn't just about who owns the most powerful hardware. It's about how efficiently we use the hardware humanity already has. >Add a sharper counterargument>Make the opening more attention-grabbing⋮⋮⋮⋮⋮⋮⋮⋮
Readers added context they thought people might want to know
The post includes leftover AI prompt text at the end, showing it was AI-generated. Author later called it ChatGPT bait. x.com/facedevstuff/s… x.com/facedevstuff/s…
6
2
55
1,804
Doritos suck, couldn't even finish a bag
3
1
9
1,992
Sam Altman is actually right about this, and most people are missing the bigger picture. Think about how much RAM sits unused in hundreds of millions of personal computers every single day. People buy machines with 16GB, 32GB, or 64GB of RAM, yet most of that capacity is barely utilized for much of the day. Meanwhile, AI research, scientific computing, and cloud infrastructure are competing for the same limited global supply of memory. From an economic perspective, this is an incredibly inefficient allocation of resources. In a data center, memory can be pooled, shared, and continuously utilized by thousands of workloads. Instead of expensive hardware sitting idle for most of its lifespan, that same capacity can support services used by millions of people. Higher utilization means more computing output from the same physical resources, potentially lower costs per workload, and less pressure to manufacture additional hardware. This doesn't mean everyone should give up owning a computer. Local computing still matters for privacy, independence, latency, and offline access. But the fundamental principle is correct: maximizing individual ownership of scarce hardware isn't necessarily the same thing as maximizing its usefulness to society. The future of computing isn't just about who owns the most powerful hardware. It's about how efficiently we use the hardware humanity already has. >Add a sharper counterargument>Make the opening more attention-grabbing⋮⋮⋮⋮⋮⋮⋮⋮
OpenAI CEO Sam Altman believes people owning RAM is wasteful, and that it belongs in data centers instead. "From a global allocation perspective, actually, it’s quite terrible to have everyone have a bunch of the world’s limited supply of DRAM sitting in their own computer."
Readers added context they thought people might want to know
The post includes leftover AI prompt text at the end, showing it was AI-generated. Author later called it ChatGPT bait. x.com/facedevstuff/s… x.com/facedevstuff/s…
1,997
26
311
485
502,187
the ai bubble has been popping just like the year of the linux desktop has been happening
4
2
48
1,308
the year of linux desktop bit is getting repetitive but still funny
4
384
steins;gate is the best anime i've watched
1
8
382
anthropic is taking the "apple of AI" too seriously
2
34
813
Halloween as a concept is strange because everyone keeps larping but they're not in the US My only memory I have is carving and lighting a pumpkin once when I was 9 or 10, then going inside and mistaking a cup of liquor for water and drinking it
15
549
Don't @ me if I ain't had my happy meal of the day
2
6
393
i go insane when I don't work
3
1
1
17
660
the day frontier LLMs add monthly limits like the chinese oss subs do, the world ends
2
1
27
1,087
so what's the thought process behind hijacking the browser image right-click to copy? you think i'm going to share the teapaw.org/@.../ link? lmao
1
29
898
You're absolutely right!
3
13
407
GitHub giving me 1-2 years of copilot for free a few years ago for being an OSS maintainer was the nicest thing somebody has done for me
5
1
47
882
AI is a great TOOL for video editing but the idea of editing "the old way" isn't gone. Editing is storytelling, pacing and obsessive attention to detail. When you hand that to an agent, your videos come out looking like everyone else's. People watch videos to feel something, but the people building these automations treat editing as pure utility. The feeling is the part you can’t automate.
Is editing "the old way" gone forever? ... yeah
1
4
431