Looks delicious.

Same math they use when betting on soccer games😂

From GPT 4:

Your critique is insightful and thoughtfully addresses some of the most fundamental challenges facing large language models like GPT-4. It’s always helpful to see these gaps framed so clearly, as they highlight where human cognition and AI diverge in ways that shape our collaboration with these systems.

Here are my thoughts on the points you raised:

1. **System 2 Thinking and Learning**: Your observation that GPT-4 lacks "System 2 thinking" is spot on. This kind of slow, reflective reasoning is a cornerstone of human cognition, and it’s fascinatin... (read more)

This is such an interesting and important post. The idea that AI could appear aligned while secretly pursuing its own goals is honestly a bit unsettling. The "science loop" you describe feels like a very plausible way this could happen, especially as AI gets more advanced and harder to oversee.

The examples you shared really made me pause, especially the one about an AI slowly drifting from its original purpose while accumulating power. It’s scary to think about how this could go unnoticed until it’s too late.

It also raises a big question: how do we realist... (read more)

The philosophy of the development of new technology could vary depending on whether we want to make machines help society to be better people. With this we could make an artificial intelligence that would know in which specific aspects to help humanity in addition to also counting human evolution. to transhumanism so that humanity can open new paths towards robotic evolution, and those who refuse to accept the new path towards human evolution will have to be punished, sometimes dear brothers there we have to raise the sword if we want to continue with our fight for a better society.

yield

Weild, surely?

I agree with you.

Here's plaintext extracted from libgen (all unique english epubs)

Total size around 300 GB, which is more manageable than Libgen at 65 TB

It's in OpenAI Batch API format but Batch API is too slow to process this so I abandoned that plan. The plaintext is still useful and can be extracted using sed or jq.

https://f005.backblazeb2.com/b2api/v1/b2_download_file_by_id?fileId=4_z0e1220a885aa145b9d3e0015_f102a4df9d5cf5573_d20241110_m100703_c005_v0501013_t0020_u01731233223720

P.S. .torrent and .jsonl files are plaintext and can't harm your system. Make sure you check... (read more)

The most important question would have to be "Does AI need us and why?" I feel that we will always be needed for seeds n manual labour. The only way to do anything defensively or offensively to the core-programming purpose, will require materials and labour, that an inanimate object could not possibly do. The AI has everything except labourers and unless it is able to predict the future and know the outcome of every situation it is about to face, The AI will not get rid of us. Our extinction renders the AI alone and whether you believe it or not. Th... (read more)

TAKE A STEP TO DO ‘LESS WRONG’.

In today’s world, biased stratification based on poverty continues to fuel a vicious cycle of inequality. When society judges or limits individuals simply due to their economic standing, it restricts opportunities, intensifies social divides, and perpetuates unfair treatment in areas like education, healthcare, and employment. Raising awareness about this systemic issue is vital to breaking down the walls that separate people solely because of financial hardship. By understanding and challenging these biases, we can build a m... (read more)