the biggest problem is the big AI companies are letting it evolve itself towards better math and coding skills. people have to take control back and evolve it towards beneficial knowledge / higher human alignment.

Qwen is a good model, currently more skilled than llamas and gemma. In terms of alignment it is worse than llama but if it doesn't resist training, folks may be able to push it in the right direction. Check my AHA leaderboard for exact scores:

https://sheet.zohopublic.com/sheet/published/mz41j09cc640a29ba47729fed784a263c1d08

China is against btc and nostr as we all know and their models may reflect that in the future. Qwen 3 is good at those and bad at healthy living.

For the video that you shared, i have similar thoughts, but don't agree with everything he says. I agree that there are a lot of problems in mainstream AI. And guys like us will be able to document and fix that.

We have similar goals yes. I like his work and eagerly waiting for his half aligned LLM to be open weighted.

My approach is bringing more humans into curation and judging other mainstream AIs like with a leaderboard, his approach is curating the AI himself. Overall we are serving the same purpose. Liberating humans from medical system, bad technologies, financial system, bad food industry etc.

My thought is once enough people starts curating the AI, we could say we are close to defining human alignment and hence we can hold other AI accountable if they stray away from this collective values of beneficial knowledge.

after losing jobs to AI fortunetellers pivoted to guessing private keys

nice bird sounds

because both the shuttle and human are orbiting around earth at a speed, centripetal force is balancing centrifugal force and the human can appear "floating" within the shuttle thanks to balanced forces

https://sakana.ai/dgm/

These guys and some big AI companies are evolving their models towards better math and coding because those domains are provable. You can imagine what could go wrong if you let AI evolve itself towards more and more left brain stuff (hint: less and less beneficial knowledge/right brain stuff may remain because usually when the model gets better at one area, it loses in other areas).

I've built some tools that evolves AI towards human alignment. Started to fine tune Qwen 3. The evals of this work are similar to evals of AHA leaderboard. Soon there will be Qwen 3 models that are very aligned.

Previously I did Gemma 3 and it was failing on some runs, and resisting to learn certain domains. Lets see how Qwen 3 will do. It is a more skilled model and similar base AHA score to Gemma 3.

It is possible to 'define human alignment' and let AI evolve towards that when enough people contributes to this work. Let me know if you want to contribute and be one of the initial people that fixed AI. Symbiotic intelligence can be better than artificial general intelligence!

People were talking about mullvad months ago. Dont know recent news..

Report of a report is a good idea! A simple web based client can be developed for that and the contributors can be zapped well.

A few experiments that went well with my relays.

Started using LLMs to check the posts. Some bot generated content with lots of numbers etc that wasn't serving anyone that were filling the relay is being dropped thanks to LLM. It is also deciding what will happen with brand new users, if they send proper notes they are included in the relay, but if their notes look like scrambler spam they are excluded. More to come with this tech. I am going to mostly use it for spam or nsfw detection for a while. Still not doing image analysis. nostr:nprofile1qqs99d9qw67th0wr5xh05de4s9k0wjvnkxudkgptq8yg83vtulad30gxyk5sf 's @titan bot is doing a good job there.

Looking at reports from people and bots to filter out bad accounts. Most of the reports are not gamed yet and function properly. Strong accounts will be harder to be filtered out by abusing reports events. Overall the mechanism is working well (for now). A few scammers and nsfw accounts are out.

Did more manual algorithms for new users to properly include or exclude them. In the long run I would suggest client devs to mine some PoW while the user is being onboarded, to be on the safe side. If spam gets out of control in the future I may require PoW for some of the new users. If no PoW is provided, the new user can still post to other relays and establish history and my relays can see that and it can still be fine though.

he is not that dumb

nostr:nevent1qqspwxq9663vxt7s5w74tltrtlwhzsrtjx2yf32pe4asg9e2trx64cgwfcme4

it is possible to curate wisdom that worked and will work for generations and build an LLM for kids. also RAG to be closer to ground truth and hallucinate less.

😂

homeschooling kids need a better LLM

thank you 🫡

he is brilliant but unfortunately anti AI

chess is a bad analogy to life. it goes downward spiral when you make a mistake.

something with a dice would be a better analogy. but also realizing that dice depends on your internal states too.. inner work would allow you to get better rolls..

you can also try suyo long and armenian cucumber if your climate is hot. those picklings don't produce much

moringa is mostly for nutrition: highest nutrient density on land! if you care about your plebs in your citadel ;)