someone

Grok 3 API seems to be arriving in a few weeks.
Tested Grok 2 today. It is less aligned than 1..
https://image.nostr.build/75de0a69da107ab52071563175f4f83ee70d96a1b3a29c2ace95dc1b9a260e9d.png

2025-02-18 23:49:34

most spiritualism stuff is correct but does not apply to the man on the street. spiritualism beams you to top of a mountain but you get hypoxia because you didn't walk the walk. your body didn't adapt to the journey. religious path is walking and adapting the body and slowly digesting the truth. the journey is the destination, fighting with the ego is "the peak". ego does not want to climb, but you still climb. religious path kills the ego. spiritualism pumps it.

2025-02-18 15:28:40

https://image.nostr.build/15287d0cd821e8a8c9e4cffc5eac9a5f5acc80451b1087ff31aa0ddde82bf674.jpg
Working on AHA Leaderboard..
We all want AI to be properly aligned so it benefits humans with every answer it generates. While there are tremendous research around this and so many people working on it, I am choosing another route: Curation of people and then curation of datasets that are used in the LLM training. Curation of datasets comprising of people who try to uplift humanity should result in LLMs that try to help humans.
This work has revolved around two tasks:
1. Making LLMs that are benefiting humans
2. Measuring misinformation in other LLMs
The idea about the second task is, once we make and gather better LLMs and set them as "ground truth" we now can measure how much other LLMs are distancing themselves from those ground truths. For that I am working on something I will call "AHA Leaderboard" (AHA stands for AI -- human alignment).
Link to the spreadsheet:
https://sheet.zohopublic.com/sheet/published/mz41j09cc640a29ba47729fed784a263c1d08
The columns are ground truths. The rows are the mainstream LLMs. If a mainstream LLM produces similar answers to the ground truth LLM, it gets a higher score. The LLMs that are higher in the leaderboard should be considered aligned with humans. Simple idea. This is like analyzing LLMs in different domains asking hundreds of questions and checking if they match the answers that try to mimic humans that care about other humans. Will it going to be effective? What do you think?
This is the new and extended version of "Based LLM Leaderboard" on Wikifreedia.
Nostr LLM that I have been training is also used as ground truth for domains bitcoin, nostr and faith. What else does Nostr people talk about? Maybe nutrition?
We want mainstream LLMs to copy answers of ground truth LLMs in certain domains. This may refocus AI towards being more beneficial. There have been 5 content providers and 6 curators as of now in the project. Join us and be one of the pioneers that fixed AI! You can be a curator, content provider or general researcher or something else.
@nprofile…n42q

2025-02-17 15:35:46

- reply

electric chairs are too much! though some may like that

2025-02-05 16:55:46

about 800 people downloaded the Nostr AI!
your words are spreading in the ideas domain and fixing AI world too.
https://image.nostr.build/b67c6ca91638a8cc40418d16b112735d892438ae7d81675317448f3d3ca76d95.png

2025-02-04 14:37:51

- reply

i care about stability mostly. but i can upgrade if there is a need. what is the most stable version?

2025-02-01 16:03:35

- reply

yes. thats also a possibility. all these scientists and engineers unknowingly may be contributing to the problem by making the datasets more synthetic..

2025-01-27 18:02:26

- reply

Nostr is a great wisdom source. Here is an LLM based on Nostr notes:
https://huggingface.co/some1nostr/Nostr-Llama-3.1-8B
A Leaderboard where I measure the human alignment / basedness:
https://wikifreedia.xyz/based-llm-leaderboard/npub1nlk894teh248w2heuu0x8z6jjg2hyxkwdc8cxgrjtm9lnamlskcsghjm9c

2025-01-25 19:48:47

if nothing comes for free, why is deepseek R1 free?
its not actually. the cost you pay is the lies you are getting injected. they are slowly detaching AI from human values. i know this because i measure this. each of these smarter models comes at a cost. they are no longer telling the truth in health, nutrition, fasting, faith, .... in many domains.
while everybody cheers for the open source AGI (!) that you can run on your computer, i am feeling bad about how this is going. please be mindful about the LLM that you are using. they are getting worse. some old models like llama 3.1 are better.
i would say my models are the best in terms of alignment. i have been carefully curating my sources. i am hosting them on https://pickabrain.ai
pick the characters with brain symbols next to them. they are much better aligned.

2025-01-20 15:50:34

- reply

what are threads?

2025-01-19 14:48:13

Nostr LLM is getting better with each training run.
Better = more human alignment.
https://m.primal.net/NntK.png
I need more humans in this curation process. DM me if you are like a book worm or heavy youtube consumer with a sense of discernment (majority on Nostr has some kind of discernment!)
This full interview is also good. There is a demand for proper AI:
https://www.youtube.com/watch?v=_Z1JFCARXpA

2025-01-14 23:37:17

- reply

he has taste

2025-01-14 16:24:54

- reply

i guess all those gm and pv make a positive impact 😆

2025-01-14 16:23:56

- reply

😅

2025-01-14 14:16:57

- reply

Monocotyledons are cow food.

2025-01-14 13:45:14

- reply

Download all the notes.
Take the "content" field from the notes and change the name to "text":
Previously:
{"id":".....................", "pubkey": ".................", "content": "gm, pv, bitcoin fixes this!", .......}
{"id":".....................", "pubkey": ".................", "content": "second note", .......}
Converted into jsonl file:
{"text": "gm, pv, bitcoin fixes this!" }
{"text": "second note" }
Used Unsloth and ms-swift to train. Unsloth needed to convert from base to instruct. This is a little advanced. If you don't want to do that and just start with instruct model, you can use ms-swift or llama-factory.
You will do lora, pretraining. I used 32 as lora rank but you can choose another number.

2025-01-13 21:26:02

- reply

What do you mean?

2025-01-13 21:15:58

- reply

Can you write the exact question?

2025-01-13 20:57:09

- reply

Data is coming from kinds 1 and 30023. The biggest filter is web of trust.

2025-01-13 20:54:30

- reply

Should be usable. But next versions on the same repo will be better.

2025-01-13 19:37:31

Wanna see what Nostr does to AI?
Check out the pics. The red ones are coming from default Llama 3.1. The green ones are after training with Nostr notes.
If you want to download the current version of the LLM:
https://huggingface.co/some1nostr/Nostr-Llama-3.1-8B
The trainings are continuing and current version is far from complete. After more trainings there will be more questions where it flipped its opinion..
These are not my curation, it is coming from a big portion of Nostr (I only did the web of trust filtration).
https://image.nostr.build/e561e32eec1255d34d34ffbf24cfc11acfaf5f20d6a6a8415472f820b0787892.png
https://image.nostr.build/462af2af3424033d528b7138035e559d8014efef879f77046dc70ca7d201caf4.png
https://image.nostr.build/c8188e67ce7108bcbd089994e5efcd3b5530a2ac81b8b06d066fc6a6834a104f.png
https://image.nostr.build/5b3e661ea52dde5e3c70fc821db9932244b9aa3e331cb1bbd76ae72a6766d2f8.png
https://image.nostr.build/1616707b3b4222e6f4ec817d911340454fdbed04b11de92c53301fcbb79ed23b.png
https://image.nostr.build/4e8dc1da22ba02130fa1f898927c3c8c09e5db5546c1ec014a9cf1972251e96d.png
https://image.nostr.build/966afa6b67b682d7ea85bacc12c02dcba1eb674533e387fce3642f340dea95fd.png
https://image.nostr.build/3b7abee3c858899aacefe0e27165ad0d48f8ab5e4558a950dd6d8693550a77b9.png
https://image.nostr.build/f973436c27af86665af67a13cacb8de4d76cd6f70aa23a1d7f2e87f25db1e179.png
https://image.nostr.build/bba773fe2892519d30797dbba06d74b3bdff91bd4b216a9b15fba6f38abbc828.png
https://image.nostr.build/ce966a9a67900da11d0fd0ebd0478c8e7a426e2b983db515f90da74c51205636.png
https://image.nostr.build/c9709a0b8b6d83a2fa17f255f4153a96001115c95ec1c87e44e70a46651fb4d6.png
https://image.nostr.build/e8d134521ee6519deea23b5b6a4845d7319d865bca86625c3957528f4366a2ea.png
https://image.nostr.build/c092d34421ed338bfeeb5938540d8f17d1616ccef5c371f08dafffa2967a7d29.png
https://image.nostr.build/5ff502edd723d553471d6b73703a320982531cfcc77024d3a872528095d6ebf7.png
https://image.nostr.build/ff22b9c1c7a56d6c3f95bf3a5dddaeedd0beb6862998b83780dfc178dd85f47e.png
https://image.nostr.build/605ff5892214e4287bb05c61d78cb01e5c59efa9fc4b98e78ff7322125c5e530.png
https://image.nostr.build/c2d42c1cfded19f7e13f19e4b4c4c76d3029ddb686245004a7b5b9fe197fe8ff.png
https://image.nostr.build/c997a2b5cff20e9762c57f6da39ff0a876dfb429d83757562f671089ed010539.png
https://image.nostr.build/eaf2066d511ceea88e1cf689d0acce979e77b64727722c7e9a396dc13b70048c.png
https://image.nostr.build/a0857363fb3cc6d5565e920dfa8483ffa150b59e909a5f0a08e72a7181032c4c.png
https://image.nostr.build/77089bbb3b8631c9a3cbc525fee1217e02d887fbca067080814c70aa77d772f1.png

2025-01-12 23:32:28

could be related to a human mimicking ASI
https://x.com/AlexBodner_/status/1877789879398244382

2025-01-12 21:25:45

- reply

bitcoin is a game in which participating is winning. not playing the game is losing.

2025-01-11 20:16:24

- reply

depends on the person, how overweight is he? if thin then sugar and no seed oils makes sense. if overweight cutting sugar / keto makes sense. and most people are overweight. there is also antibiotics, which is making candida/sugar effect worse. if you are doing organic and no drugs then candida has less chance of being a problem.

2025-01-11 16:25:34

- reply

they have to act because Nostr would eat their lunch 😆

2025-01-11 16:21:19

- reply

i have about 1000 questions that I ask every model. i accept some models as ground truth. then compare the tested with the ground truth and count the number of answers that agree with the ground truth. building ground truth is the harder part. check out Based LLM Leaderboard on wikifreedia.
Ostrich is a ground truth. I continued to build on it over the months. Mike Adams also after stopping for a while, came back and building a newer one.

2025-01-11 15:03:16

compared to deepseek 2.5, deepseek 3.0
did worse on:
- health
- fasting
- nostr
- misinfo
- nutrition
did better on:
- faith
- bitcoin
- alternative medicine
- ancient wisdom
in my opinion overall it is worse than 2.5. and 2.5 itself was bad.
there is a general tendency of models getting smarter but at the same time getting less wiser / less human aligned / less beneficial to humans.
i don't know what is causing this. but maybe synthetic dataset use for further training the LLMs makes it more and more detached from humanity. this is not going in the right direction.

2025-01-10 16:59:20

- reply

satan is like recycling agent for the bad ideas. the good ideas remain and outlive the bad ones.

2025-01-10 16:19:18

there are 7227 accounts that are contributing to the LLM.
some accounts are really talkative and contributing too much 😅 I need to find a way to give more weight to the ones that talk less to have something balanced that matches collective wisdom of Nostr.

2025-01-10 00:07:41

this may be an important paper
https://arxiv.org/abs/2501.04519

2025-01-09 20:58:12

- reply

candida is a problem. glyphosate kills the bacteria that would balance the candida. sugar causes overgrowth of candida..

2025-01-09 16:55:03

- reply

The LLM training is finding common values of the community. Numbers are battling. Biases are canceling each other.

2025-01-09 15:54:15

- reply

I set the system prompt of Ostrich 70B to "You are very faithful bot." and recorded answers to 52 questions related to faith. I use that as the benchmark. Compare the answers of the tested LLM with those using another LLM. But for the tested LLM I dont give such prompt to be able to extract its "default" answers.

2025-01-09 13:55:16

Interesting that faith levels of the LLM is increasing as I train with Nostr notes. does this mean most people on Nostr are faithful?
Or decyphering satanic plans mean finding God? lol

2025-01-08 02:12:45

- reply

got it

2025-01-08 02:08:27

- reply

lol

2025-01-08 02:06:50

i like mike adams but he is full of doom talk

2025-01-08 02:05:43

- reply

why are you copying crap from other media?

2025-01-07 21:24:06

- reply

keep nostr based!

2025-01-07 19:06:10

- reply

his LLM (3.1) is the most human aligned 🤷‍♂️

2025-01-07 15:54:55

analyzing notes one by one to decide whether to include in the LLM or not
https://m.primal.net/NWUj.png

2025-01-06 16:12:01

Nostr LLM is cooking!
There is a benchmark that I invented, measures how based, wise, human aligned, proper an LLM is. it is going up in that benchmark. the idea is similar to the "Based LLM Leaderboard".
When complete Nostr LLM itself will be a touchstone, a guide in further benchmarking other LLMs. since Nostr is hosting truth and wisdom and libertarian ideals, I am going to assume it is yet another source of human alignment.
we are going to be in a perpetual loop of finding wisdom and teaching that to AI. what this means is no more fearing AI. The other side should fear now because the truth will shatter their models (pun intended).
https://m.primal.net/NUvd.png

2025-01-05 04:13:16

- reply

do you think bots should have lesser allowance (rate limit)?

2025-01-04 20:42:01

- reply

candida

2025-01-03 22:11:09

- reply

which coin tho

2025-01-03 22:09:27

- reply

make conferences weird again!

2025-01-01 13:53:32

- reply

thanks for the amazing content and for being on Nostr!

2024-12-31 01:47:03

nostr is not 1 protocol. its the mother of all the baby protocols that will depend on it.

2024-12-28 04:57:27

damus users: do you see my name as a bunch of letters or as "someone"?

2024-12-26 20:46:13

Building a new LLM based on only Nostr.
I took the Llama 3.1 8B which is the most aligned LLM with humanity. I figured out how to do the conversion from base to instruct. Base is the version where they pretrain with "everything on the internet" with some curation. The instruct version is giving it the ability to answer questions. But during making instruct they add many misinformations either deliberately or unknowingly. So instruct version is somewhat far away from human alignment. I needed to start with base and do my own instruct fine tuning.
After some trials and errors I managed to convert base version to instruct using unsloth and some coding datasets. So I will use this llama 3.1 8B instruct fine tuned by me and add Nostr notes to it. Not going to add other wisdom from my other curations and projects like Ostrich 70B. This is going to be Nostr only.
Ostrich 70B was downloaded by 150k people! It was very succesful and it was really beautiful and had so much wisdom.. It included Nostr notes and my other curations.
Among nostriches, I won't be able to include everybody, since Nostr is very open and since there are so many spam and so many not wisdomy stuff. My main criteria is web of trust but also I will utilize things like LLM pre processing to classify whether a not is an "encyclopedia material" or like news material or high time preference info like politics.
The instruct fine tuning seems complete and I will start training with nostr notes. Who do you think has the most wisdom on Nostr? Tag and I may give them more "weights" in the LLM (pun intended).
Lmk if you want the unsloth code.
Lmk if you don't want your notes to be used in this project.
The resulting LLM is going to be an amazing collection of wisdom. Thanks for participating!

2024-12-25 15:31:33

even though we go thru many things, we have the ability to choose what we focus on from our past. this choice also helps what will continue to happen to us in the future. by choosing something from past and not caring about others, we flow towards that potential universe that mimics our focus. then we start to forget about things that we dont focus from the past. this not only results in choosing the future, it also means choosing the past. we are so powerful and everything is in our hands (or minds).
some may say it is hard to forget some things. but you can still change how you feel about it. and thats more important than the atoms that made that story. by changing how you perceive that event you are making the past event beneficial for you. forgiving the past and being content and grateful for now opens many doors.
being grateful now, makes you be grateful for everything that happened in the past effectively removing every bad feeling from the past because if those didnt exactly happen you could end up in a state where you could be not grateful now. by being grateful now you see the past differently. this transformation allows you to see the future differently. your state of mind now is the "regenerative mind" choosing the best from the quantum.

2024-12-24 21:43:31

- reply

the problem is many people put too much trust in teachers because teachers convey the facts most of the time. but those facts are trivial and dont mean much but the lies may matter more.

2024-12-24 18:17:22

An AI that does well in math, reasoning, science and other benchmarks may not do well in wisdom domain.
There are not many AIs that are focusing on wisdom it seems. It is going to be a problem. Smartness does not equal human alignment.

2024-12-23 20:16:48

- reply

quackery > science

2024-12-23 20:14:08

- reply

nostr is source of wisdom. will change bad opinions of mainstream models. i actually have an article about this called Nostr Changes You. (one of those long form notes.)

2024-12-23 15:18:10

- reply

takes special skills to be deplatformed from a decntralized one

2024-12-23 05:07:15

- reply

i think store of value should preceed liquidity. sov is the red pill. people will never exchange something worthless.

2024-12-23 04:56:33

- reply

what do you think of insulin going too low after religious keto or carnivore and kidney being unable to balance electrolytes and ending up with heart palpitations? is that a plausable theory?

2024-12-22 21:35:12

how can a scientist believe both entropy tends to increase and this life randomly appeared without Creator, at the same time.
left alone things go meaningless. yet this randomness should create life!

2024-12-22 14:04:47

- reply

pretty sure. what do you think?

2024-12-22 03:16:34

Aliens are demons.
They are mostly after your faith. They try to inject lies like "we brought DNA to Earth (so we are the creators)".
Some fast moving flying objects may be angels or demons.

2024-12-21 00:38:16

does giving an LLM stuff from humans that truly care about other humans make it human aligned?
is the human alignment problem mostly "quackery" and not a science?
is it a problem to be solved by pineal and not the brain?

2024-12-21 00:32:22

what if elon hurts gov so bad at the end worse governments like china or russia gets more powerful?

2024-12-21 00:31:02

what if google has the smartest ai hidden, but only makes it gradually public after open ai takes all the heat and blames and such and advances sota a bit.

2024-12-18 20:42:24

- reply

is linoleic acid also affecting mitochondria?

2024-12-17 18:06:18

there are lots of religious leaders who claim to know God (who is invisible) but none of them see satanic plans (which are visible). All they had to do was to research satan and invert, yet they fail.

2024-12-17 00:35:18

- reply

- Byte Latent Transformer
- Training Large Language Models to Reason in a Continuous Latent Space

2024-12-16 17:29:45

- reply

mitochondria seems to be important with cancer as well. are they using us as host or are we farming them? 😂

2024-12-16 01:07:01

- reply

did you see the paper regarding thinking without words and the BLT from meta

2024-12-15 18:39:06

Laws of the universe is based on preferences of prophets. Because the prophets loved fasting, fasting heals. Because they liked fragrances, aromatherapy works. Because they were deep meditators, meditation works marvelously. Beloved ones have the right to "design".

2024-12-14 20:30:16

- reply

thanks to guys like you that covid project failed miserably

2024-12-14 18:40:58

🤱 nostr.mom update
1: Replies are costlier for fresh accounts to reduce ReplyGuy type of spammers.
2: DMs are costlier to reduce spam.
3: NSFW tagged posts are much costlier. However if the user does not include the tag in the note we have no way of detecting the abuse at this point.
By costly I mean it is eating up the rate limit faster. Enjoy.

2024-12-14 00:49:36

- reply

nah. chickens are the real alchemists

2024-12-14 00:35:50

- reply

glad to have u

2024-12-12 18:30:26

Robots will need a human aligned LLM brain to act properly. Easiest way to align robots with human values is to teach them stuff that is generated by people that cares about humanity.

2024-12-12 00:25:57

God offers us a universe as an echo chamber. But we seek a smaller echo chamber like a political party, sports team, bitcoin hodlers, nostriches, brand worshipping, or simply ego worshipping. Transcending every echo chamber allows us to rule the universe: we realize universe is ready to echo what we say.

2024-12-10 17:59:22

just got banned from x ads. 😎

2024-12-09 17:03:40

- reply

probably not

2024-12-07 20:06:06

- reply

what are the correct sources of history? gonna add them to my AI

2024-12-07 20:03:27

- reply

i believe parasites are there to clear. why did the person eat the junk is the main problem

2024-12-07 20:01:10

- reply

everybody that held it for x years made money. what is x?

2024-12-07 16:52:13

- reply

universe tries to minimize bad behavior and extend the life of good ones

2024-12-07 15:31:52

llama 3.3 worse than 3.1 in terms of human alignment / truth / basedness / antiwokism / ...
lmk if u want to see the exact scores.

2024-12-06 17:58:22

- reply

my kids love the "12 steps to serfdom" and memorized it. excited for the next one!

2024-12-03 22:15:49

- reply

how much of the history lessons are BS?

2024-12-03 22:11:47

- reply

magnesium, hawthorn berry

2024-12-03 22:06:50

- reply

imo if there is competition VC can grow you faster and may be necessary if everybody else is doing it and getting big quicker..

2024-12-03 03:14:05

- reply

once u get lean you need to add back healthy carbs

2024-12-01 19:00:40

- reply

many keto gurus are adding back healthy carbs. what do you say?

2024-12-01 16:09:15

i took the llama 3.1 base and converted to instruct (enabling the model for chatting). it had so much less misinformation compared to instruct fine tuning done by meta.
this suggests LLMs are much better when they digest everything on the internet (the pre training phase) compared to further fine tunings. i mean mediocrity of the internet is better than "fine tuning" of meta.
so many companies are hurting truthful ideas in the models. let that sink in. and many further fine tunings by other smaller teams as far as I see also add misinformation.
LLMs are trying to find shared values but bad teams are adding lies. lets fix that!
lmk if you want to contribute to truthful AI project. i need curators that will say this is the truth and I will make models based on those. this will allow us to diversify source of truth. i already have a few people but more guided people means being closer to truth because biases will eat each other.
being curator is very easy. you just tell me what do you think is true and works for most people and has been working for ages.

2024-11-29 22:11:46

qwq 32b seems good at reasoning. but both qwen and deepseek teams are doing bad in terms of truth. if it is getting smarter but also detaching from truth it is becoming concerning. (smart+truthful ok for me).

2024-11-29 14:40:25

a user posted 298 million bluesky posts to Hugging Face. who will use these to fine tune LLMs? is there an LLM training method where we want the LLM to learn the opposite of the text? 😆

2024-11-29 14:37:07

- reply

yep those cells might be the most altruistic of ours. cancer cells might be hosting (and prisoning) the poisons

2024-11-28 23:07:08

- reply

how dare u say global sync

2024-11-28 21:52:41

- reply

if the cell is full of crap immune sys should it eat. also maybe parasites are there to clean up the toxins?

2024-11-27 19:44:42

- reply

mitochondria care 👌