harhargange

6 ms·

1.
by harhargange 7d ago
I’m pretty sure that OpenAI has some of the best mathematicians prompting the models and analysing the results. While they are marketing as if the model solves problems themselves.
2.
by harhargange 7d ago
OpenAI has messed up big time here by competing with their customers. It would have become the norm for humans and mathematicians to use the tools and publish bigger results any way. If OpenAI didn’t run for credit, this theorem itself may
3.
by harhargange 8d ago
Not the hope, but telling them that an answer does exist. This ofcourse means now we are going to go into an even more darker cave next time we are looking for some gold and pathbreaking discoveries will become even more rare. Add to that,
4.
by harhargange 8d ago
This is definitely true. So i can now justify my recent gpu purchase
5.
by harhargange 8d ago
I have a dumb feeling that the proof will be wrong with serious flaws but that will be found out only after the ipo
6.
by harhargange 8d ago
They are basically playing with the dates so that they can claim their results 'accidentally' got trained when they were training the new model.
7.
by harhargange 8d ago
Check this https://cims.nyu.edu/~tristanb/statement.pdf
8.
by harhargange 8d ago
Just so everyone knows, although openAI pretends that the model generated solution and wrote the paper by itself ""with very little human input"" as Buckmaster himself mentioned in his statement. In reality they have tea
9.
by harhargange 8d ago
This should be on the front page
10.
by harhargange 9d ago
“I asked when the first prompt had been sent by them. This question was not answered directly by OpenAI for some time. Eventually it was agreed that it had been sent in the past few days, after information about our work had reached OpenAI.
11.
by harhargange 11d ago
I recently bought a motherboard for my first self made pc. Six to eight sata slots is already enormous and the fact that you pay once and use your hardware freely opened my eyes. I was paying a subscription for cloud 1 tb for a trivial serv
12.
Softwares that are not tools(harssh.substack.com)
2 points by harhargange 18d ago|0 comments
13.
by harhargange 19d ago
This is also why i like the blackberry still. iPhone uses too much of cognitive overhead. At least the iPad is a good mix and i mostly depends on my keyboard/mouse connected iPad to use my apps if I’m not virtualising them.
14.
by harhargange 26d ago
Having done all sorts of experiments with android in the past, i now prefer not touching my devices and getting into a rabbit hole. The optimisations never end and you get confused what you’re meant to do on your device. In this sense Apple
15.
by harhargange 29d ago
So you can still use the hardware via software fix?
16.
by harhargange 29d ago
I recently built a pc with 5950x and 7900xt. What do you mean by diminishing returns?
17.
by harhargange 1mo ago
I had been planning to buy a GPU for Blender to compliment my 5950x CPU. The moment I ran the Qwen3.6-27b on my CPU, I arrived at the conclusion about the GPU I want. I saved some money and ordered the 7900xt-20gb for around 600 USD (instea
18.
by harhargange 1mo ago
Along with remembering facts, it is important to insightfully give ‘weight’ to the facts. AI gives ‘weights’ to things depending on its own priorities or corporate interests. Our brain gives the weights to various facts depending on persona
19.
by harhargange 1mo ago
I too purchased the 7900xt as it was cheap with a lot of vram. Qwen 3.6 27b gives me 30 tok/s
20.
by harhargange 1mo ago
Physics laws are the ultimate form of compression because they are so universal and say so much about so many things in few words, or a formula. This is why Newton's laws were a big achievement. And they enabled predicting the behavior
21.
AI on a Piece of Sponge?(driftspeed.wordpress.com)
1 points by harhargange 1mo ago|1 comments
22.
by harhargange 1mo ago
Sorry if you find this self-promoting, but I wanted to get this piece out and share it with the community. This is some of the research I do. Feedback is welcome.
23.
by harhargange 1mo ago
Exactly, what will be the size of big models? Maybe they aren’t targeting big models but where do they give size estimates?
24.
by harhargange 1mo ago
I have an AMD CPU as well as an AMD GPU. Their open-source philosophy and Linux compatibility, including a clean core distribution (No big.little for desktop) is what keeps me in. I think Taalas is in safe hands.
25.
by harhargange 1mo ago
My 7900XT runs it at 35 tokens per second.
26.
by harhargange 1mo ago
The 27b model inference speed is amazing at 35 tok/s with power limit to 260W. My cpu is 5950x, maybe that matters too. Really enjoying it and it’s my default now.
27.
by harhargange 1mo ago
I used the qwen 3.6 models and convinced myself to go ahead and get a gpu. I got the 7900 xt with 20 gb vram.
28.
by harhargange 3mo ago
I was building a pc and the ai was hellbent on saying either the cpu pins are bent or motherboard is faulty. Both giving me a panic attack. The solution came from updating bios which i had done so ai ruled it out. But the shop owner did the
29.
I Block All Browsers(substack.com)
1 points by harhargange 6mo ago|0 comments
30.
by harhargange 9mo ago
And the bb10os was based on this. I still use it and it’s the ultimate phone that supports few android apps still.

More ›