🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

o3 generates a fix, and this ChatGPT split screen appears (Response 1/Response 2 and which one do you prefer question.) In the fix on the right: "Add two new line characters in this one place." In the fix on the left: "Fix these 3 files in these 3 places." Which one do I

LLMs haven't reached the level of autonomy so that they can be trusted with an entire profession, and it's already clear to everyone except for ignorant, deranged, or delusional people that they won't reach this level of autonomy. What remains is LLMs as a tool that humans use.

Oh, a consensus among the deranged ones has emerged 🎉🥳🪩

Illiterate people don't know where to use em dash. When they see such a symbol in a text, they automatically assume the text was AI-generated. They can't think that there are people loving English and following its rules.

The Hundred-Page Language Models Book by Andriy Burkov is on sale on Leanpub! Its suggested price is $50.00; get it for $20.00 with this coupon: leanpub.com/sh/e1ZpBUtn Andriy Burkov #Ai #Gpt #NeuralNetworks #DeepLearning #DataScience #ComputerScience #Linguistics

The Hundred-Page Machine Learning Book (PDF + EPUB + extra PDF formats) by Andriy Burkov is on sale on Leanpub! Its suggested price is $40.00; get it for $14.00 with this coupon: leanpub.com/sh/G87I4WkU Andriy Burkov #DataScience #ComputerScience #MachineLearning #Ai

I might be wrong but I think that Google doesn't release a 70B-parameter Gemma because it would cannibalize Gemini.

😍

Since Trump was elected, Americans (and the rest of the world) have learned how they have been robbed by making them pay high taxes and using them to pay for some doubtful (to say the least) campaigns abroad. One might think maybe these abroad campaigns were needed to maintain

Each time I post something about Trump, a bunch of idiots unfollow me. I consider this technique "cleansing by Trump." These people hate Trump so much that even when he reveals how they have been abused and lied to by the government for decades, they will still hate him and want

While I wholeheartedly hate StackOverflow, reading the answers for 10 minutes on StackOverflow is reading something for 10 minutes that worked for someone. When I read for 10 minutes an explanation of a solution generated by an LLM, it's very much likely reading for 10 minutes

This made me think that I can define a benchmark for how well a model recommends music. I have about 500 songs in my liked songs. I use a random 200 to pull recommendations from the model. I can then compare the overlap of recommendations with the remaining 300 that the model

Llama 4 Maverik doesn't even deserve to be talked about in 2025. It's so uncouth and homely, and dumb. I feel the 2023 vibes that I started to forget the feeling of.

While Gemini is quite verbose, its explanation skill is the best right now. It codes best from scratch too. To edit existing code, I ask Gemini to explain what to do and then feed the explanation to Claude. I have absolutely no use cases for ChatGPT. However, ChatGPT beats

That's funny that in the era of Photoshop skills now available to a schoolboy, they prove a point with a counterphoto. I also have a counterphoto that counterproves their counterproof.

Study history. It has all the answers👍

Poor reinforcement learning researchers. They first invent a nice, intuitively motivated and mathematically sound learning algorithm, but they, unlike their supervised learning cousins, don't have a powerful computational engine, like PyTorch, built specifically for their