That's it. I QUIT! No Asian loving for a loooong time!

Member Since
Apr 16, 2026
Posts
1,148
Likes Received
3,042
Location
Virginia, USA
I finally pulled the trigger early this morning. For a long time I had been torn between buying peps from sketchy Chinese peddlers and the other thing. But things came to a head a few hours ago. I saw this video of Dr.Alex Tatem on Diary of a CEO, and almost went down the crypto abyss. Then the Tony Robbins-Igor AI video came into my feed, and I got an epiphany. I needed to buy an AI laptop, but I had been prevaricating for months. Then the memorial day discount arrived and I finally succumbed. The massive investment means that I won't be able to purchase peptides anytime soon. It is a shame, but I'm broke. And I mean real broke. I will now have to chase Ben Franklin instead of driving out lard of the future. Got to work on the business.

But, it will mean that I will be more "regular" in posting here. Yeah, my current laptop had several keys malfunctioning, especially the letter P, and the numbers seven, eight, and nine. It was such a pain copy-pasting them. After tomorrow, my typing speed will return to baseline levels.

P.S:-I know there are a few folks here who have issues with my rapid posting. But I also know that there are ladies here. So the only appropriate response I have for you folks is what the T-800 had for the motel sup who asked him "Hey buddy, you got a dead cat in there or what?" in 1984.
 
I can't imagine what I'd use an "AI laptop" for, although my imagination would probably be stoked if I knew what it was.

I'm not stranger to spending money on consumer electronics and using AI, but a quick Google search shows nothing particularly interesting. I have most AI features turned off on my laptops and phone, they aren't generally helpful for anything I do.

What I do, typically uses variety of AI or machine learned programing to increase workflow.
 
I can't imagine what I'd use an "AI laptop" for, although my imagination would probably be stoked if I knew what it was
A major drawback for folks like me who want to use AI for business purposes is that we have to use public AI LLMs with public cloud storage. Running your own personal, local LLM is a dream for many of us, but for that, you need an AI Laptop that has enough juice to make it practical. That's what I bought now. The laptop will let me run my own LLM, train it with info only I can provide, and most importantly, keep the proprietary info private and secure.
I have most AI features turned off on my laptops and phone, they aren't generally helpful for anything I do.
Ah, thats AI enabled features. I was talking about running my own LLM on my computer with a 800 Tops processing power.
Is it even possible? 😆
Oh ye of little faith! tsk tsk... Kids these days have no idea of what's possible.
Cate Blanchett Ragnarok GIF by Marvel Studios
 
A major drawback for folks like me who want to use AI for business purposes is that we have to use public AI LLMs with public cloud storage. Running your own personal, local LLM is a dream for many of us, but for that, you need an AI Laptop that has enough juice to make it practical. That's what I bought now. The laptop will let me run my own LLM, train it with info only I can provide, and most importantly, keep the proprietary info private and secure.

Ah, thats AI enabled features. I was talking about running my own LLM on my computer with a 800 Tops processing power.

I assumed you were talking about running your own AI/LLM, but if I was doing that I don't think I'd want it running on a laptop.... I'd put it in the network closet and not pay the premium to have it in a laptop form factor, deal with the power consumption/thermals, and be limited by presumptively only accessing it from the laptop vs network access from multiple clients.


~800TOPS is a single RTX5090'esque performance, right?
 
I'd put it in the network closet and not pay the premium to have it in a laptop form factor,
Yes to almost all of it. And yeah, I have spacetimecash issue. I am getting some extra hard-drives and yes eight hundred TOPs is from a single Geforce RT5090 chip. I am using the laptop for the portability I will need for my creative endeavor where I will have to brainstorm with the AI when ideas pop into my head. I am a writer you see. And all the post-writing here gives me practice--writing juice--. Thanks a ton for the info.
 
Last edited:
A major drawback for folks like me who want to use AI for business purposes is that we have to use public AI LLMs with public cloud storage. Running your own personal, local LLM is a dream for many of us, but for that, you need an AI Laptop that has enough juice to make it practical. That's what I bought now. The laptop will let me run my own LLM, train it with info only I can provide, and most importantly, keep the proprietary info private and secure.
Obviously this is unrelated to main purpose of this forum but just something I would note about this approach is there are some issues you will certainly encounter with this.
  1. Training an LLM, especially on your own computer, is extremely time-consuming (on the order of days, weeks, maybe longer.
  2. Using only info that you provide will probably bias the LLM towards whatever information you cherry pick to train it on.
  3. The LLM you make will be significantly weaker/worse than the latest models from ChatGPT and the like.
A stronger laptop is not a bad thing at all though, might I recommend instead downloading a solid existing LLM? Check out ollama dot com to start they have some good models. Also all of the info will remain private on your computer this way (like how when you download and install Microsoft Word, nobody can see what docs you make unless you upload them to the cloud or put them on a server somewhere).
 
might I recommend instead downloading a solid existing LLM? Check out ollama dot com to start they have some good models
That's the plan. I will download a base LLM model from somewhere with all the basic algorithms..Then train it with my material and then use it.
I feel like we would probably have very similar YT feeds if we compared
Hmmm...I'm not certain. I only saw Tatem yesterday. My feed is very 80's hypermasculine.
 
might I recommend instead downloading a solid existing LLM?
Local model hosting chad

A major drawback for folks like me who want to use AI for business purposes is that we have to use public AI LLMs with public cloud storage. Running your own personal, local LLM is a dream for many of us, but for that, you need an AI Laptop that has enough juice to make it practical. That's what I bought now. The laptop will let me run my own LLM, train it with info only I can provide, and most importantly, keep the proprietary info private and secure.
I work in the AI field. It was also my grad degree. Expect to have to use a tiny model. Rule of thumb is 16gb of memory per 1B parameters for training, not counting the size of the batched training data in memory.

if you just want to run inference you can exepct to need about 1GB memory per 1B params running 8bit quantized weights, or half that running 4bit quantized weights.

unless if you have a gpu in your laptop expect the tokens per second output to be modest, maybe 10-20

Im not trying to detract from what you are doing, i'm cheering you on.

dm me if you want to ask any questions
 
I work in the AI field. It was also my grad degree. Expect to have to use a tiny model. Rule of thumb is 16gb of memory per 1B parameters for training, not counting the size of the batched training data in memory.

if you just want to run inference you can exepct to need about 1GB memory per 1B params running 8bit quantized weights, or half that running 4bit quantized weights.

unless if you have a gpu in your laptop expect the tokens per second output to be modest, maybe 10-20
64GB RAM, GeForce RT5090...that's all I know.
 
64GB RAM, GeForce RT5090...that's all I know.
with gpu you are limited to using the onboard gpu vram if you train with the gpu, so maybe 24gb vram is available.


what you want to look into is training a LORA, not an entire model. that would optimize what you are trying to do.

LORAs sit on top of existing models and are much much easier to train. the large majority of people dont actually train their own models anymore, they train LORAs instead to specialize an existing model in something

LORAs use about 1/5 the memory to train as an entire model, so you can use a large model and train a LORA more effectively than using a model 1/10th the size and getting worse overall results because its a worse model
 
Based on the title I thought you were giving up visits to the massage parlor with happy endings.
I'm offended you think me so shallow.. This is so NOT the way, Grogu.
with gpu you are limited to using the onboard gpu vram if you train with the gpu, so maybe 24gb vram is available.


what you want to look into is training a LORA, not an entire model. that would optimize what you are trying to do.

LORAs sit on top of existing models and are much much easier to train. the large majority of people dont actually train their own models anymore, they train LORAs instead to specialize an existing model in something
Lora it is. I need a new sister-wife for Reta and Tesa anyway
 
I finally pulled the trigger early this morning. For a long time I had been torn between buying peps from sketchy Chinese peddlers and the other thing. But things came to a head a few hours ago. I saw this video of Dr.Alex Tatem on Diary of a CEO, and almost went down the crypto abyss. Then the Tony Robbins-Igor AI video came into my feed, and I got an epiphany. I needed to buy an AI laptop, but I had been prevaricating for months. Then the memorial day discount arrived and I finally succumbed. The massive investment means that I won't be able to purchase peptides anytime soon. It is a shame, but I'm broke. And I mean real broke. I will now have to chase Ben Franklin instead of driving out lard of the future. Got to work on the business.

But, it will mean that I will be more "regular" in posting here. Yeah, my current laptop had several keys malfunctioning, especially the letter P, and the numbers seven, eight, and nine. It was such a pain copy-pasting them. After tomorrow, my typing speed will return to baseline levels.

P.S:-I know there are a few folks here who have issues with my rapid posting. But I also know that there are ladies here. So the only appropriate response I have for you folks is what the T-800 had for the motel sup who asked him "Hey buddy, you got a dead cat in there or what?" in 1984.
View: https://x.com/AnisCodex/status/2057479106762273069?s=20
 
I'm offended you think me so shallow.. This is so NOT the way, Grogu.

It's more of a reflection of where my mind is at right now 🤣.

So, it's about me, not you 😝. No offense intended.
 
I finally pulled the trigger early this morning. For a long time I had been torn between buying peps from sketchy Chinese peddlers and the other thing. But things came to a head a few hours ago. I saw this video of Dr.Alex Tatem on Diary of a CEO, and almost went down the crypto abyss. Then the Tony Robbins-Igor AI video came into my feed, and I got an epiphany. I needed to buy an AI laptop, but I had been prevaricating for months. Then the memorial day discount arrived and I finally succumbed. The massive investment means that I won't be able to purchase peptides anytime soon. It is a shame, but I'm broke. And I mean real broke. I will now have to chase Ben Franklin instead of driving out lard of the future. Got to work on the business.

But, it will mean that I will be more "regular" in posting here. Yeah, my current laptop had several keys malfunctioning, especially the letter P, and the numbers seven, eight, and nine. It was such a pain copy-pasting them. After tomorrow, my typing speed will return to baseline levels.

P.S:-I know there are a few folks here who have issues with my rapid posting. But I also know that there are ladies here. So the only appropriate response I have for you folks is what the T-800 had for the motel sup who asked him "Hey buddy, you got a dead cat in there or what?" in 1984.
I read the title of your post and thought, who the hell would give up Asian women. But then I read the post.
 

Trending Topics

Forum Statistics

Threads
18,046
Posts
187,901
Members
60,520
Newest
retaddict
Back
Top Bottom