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.
someone
npub1nl…hjm9c
2025-01-11 15:03:16
Author Public Key
npub1nlk894teh248w2heuu0x8z6jjg2hyxkwdc8cxgrjtm9lnamlskcsghjm9cPublished at
2025-01-11 15:03:16Event JSON
{
"id": "cbebef72933547e012770d35898f44c694ce5e3782736ed68b7a9e7d0af8d3f1",
"pubkey": "9fec72d579baaa772af9e71e638b529215721ace6e0f8320725ecbf9f77f85b1",
"created_at": 1736607796,
"kind": 1,
"tags": [],
"content": "compared to deepseek 2.5, deepseek 3.0 \n\ndid worse on:\n- health\n- fasting\n- nostr\n- misinfo\n- nutrition \n\ndid better on:\n- faith\n- bitcoin\n- alternative medicine \n- ancient wisdom \n\nin my opinion overall it is worse than 2.5. and 2.5 itself was bad.\n\nthere is a general tendency of models getting smarter but at the same time getting less wiser / less human aligned / less beneficial to humans.\n\ni 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.",
"sig": "f8a30e5b13447ecdd65e00d38d7a0d43c8b16ece1d350be1050e83433fc73b37e976fd917244bb9aad1224f5f1f583e44ab7c3aaa97b7309005fe5beec59b37e"
}