(PM me) 1. You must be very serious about wanting to create AGI. You might be such a person if you have little emotional drama attached and are a hard worker during any time being put in. Motivation is essential for choosing team mates. Team members must only talk about work towards AGI; only 2% of the time can be spent on memes, cars, goofing around, etc. The goal is to act like a robotic hive that helps each other become as intelligent as each other (clones).
2. Experience. While already knowing how to code from scratch ex. "GPT" is a big reason to choose team mates, there is more reasons: Someone less experienced but that has 24/7 time 'can' do as much as an experienced with very very little time. Also, if they have solid discoveries to build on the SOTA but can't even program, this 'can' be as big a reason to choose them. It is not bad to try new team mates and then if they are not a good fit at some point to temporarily remove them. For me, I have 24/7, a few solid discoveries, and an intermediate experience. 3. Mentoring each other. If team members know other members have the prerequisites needed to join the group, they should be happy to spend lots of time mentoring each other instead of linking papers and articles for questions asked. No one wants to mentor someone that won't be fruitful, nor do they want to listen to someone inexperienced. So any time a member asks a question, we must use natural language to answer the question and not link anything, this makes learning much faster / less stress. We must also share too. It can take time to figure out how to get at a question, but brain to brain, once you do you can teach even all the details of GPT in the matter of a few days, which is rather a hard thing to do. 4. GPT, like the human brain, can predict music, image, and text nearly as well. No other algorithm I know of can do this efficiently. If you can't show me such an alternative algorithm, you MUST tell me and others you will be using GPT and building onto it to get to AGI. Trying to build your own architecture is fine too out of refusal but you must admit the SOTA has a higher probability at being good and takes time to make an alternative to GPT, so you should at least say you will give it 50% chance, not "screw GPT, I have my own ideas". You need GPT. Start showing results if you have an alternative. Start sharing. Things like this are simple, but no one can handle some critical thinking sadly it seems. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Te8844e63f5aa7b5c-M79779782fec3704ebe9faa6a Delivery options: https://agi.topicbox.com/groups/agi/subscription