raindear

There has been a slow progress in computer vision in the last ~5 years. We are still not close to human performance. This is in contrast to language understanding which has been solved - LLMs understand text on a human level (even if they have other limitations). But vision isn't solved. Foundation models struggle to segment some objects, they don't generalize to domains such as scientific images, etc. I wonder what's missing with models. We have enough data in videos. Is it compute? Is the task not informative enough? Do we need agency in 3D?

Are dead ReLUs still a pronlem today? Why not?

How far can you go by improving the curriculum? Start simple. Find a shorter and shorter sequence of examples that gives you thd best result. What is the shortest sequence to get to some perplexity? Why?

Progress is what happens thanks to AI skeptics busy defining model limitations. The limitations set attractive bars to pass.

I read that deepseek was trained on western llm output. So it is expected to have the same biases.

NYT: https://news.ycombinator.com/item?id=42403886

Molecular Biology research is much more funded than computer science. At least in Academia.

Yes, cDNA is a DNA fragment that does not interact with the patient's DNA in this treatment. Wikipedia: "Once amplified, the sequence can be cut at each end with nucleases and inserted into one of many small circular DNA sequences known as expression vectors. Such vectors allow for self-replication, inside the cells, and potentially integration in the host DNA. They typically also contain a strong promoter to drive transcription of the target cDNA into mRNA, which is then translated into protein."

But why do transformers perform better than older language models including other neural language models.

There are around 40k Hamas gunmen. So the reported numbers can get that high.

What would you do if an enemy army crossed the border and slaughtered 1000 men, women, children, party goers and elderly?

Israel withdrew from Gaza in 2005. There have not been tanks there since.

Gaza Strip borders with Egypt too.

The article sympathizes with only with one side of a horrible war, and only about losing its internet connection. It's surrealistic. It's like reading an article about Berlin citizens losing telephone connection in May 1945.

For example, you can create an embedding for an image using a neural net that has been trained to receive an image and output a vector of 1024 floats, which represent the content of the image. This vector is a lossy compressed version of the image.

A competition to self driving cars? If urban flight is possible in 10 years, will people need self driving cars?

It's not available for smartphones.

Yes, the training is based on predicting a missing word. At first, such a task allowed GPT2 to generate paragraphs that were surprisingly coherent, at the user's request. ChatGPT added a novel ability of conversation. Regarding timestamps, the training did include many examples of timestamps in JSON format.

The date of the post is April 1st.

It's not the task, it's the co-workers.

But it could be detected with RF.