Comments by ahmadmanga (@ahmadmanga) - Proof of Brain

ahmadmanga

www.binance.com

Fly with me out of Reality's boring corners to Imagination's Fascinating Horizons! (Muslim. Writer. Artist.)

about 3 hours ago

RE: LeoThread 2025-02-23 00:00 Love is a bottomless pit with burning spikes at the bottom, that's why you "fall" in it! I just realized I said, a bottomless had a bottom, but let's ignore that!

1

0

0.000 POB

about 21 hours ago

RE: LeoThread 2025-02-21 13:43 Compared to them Llama 3.3 70B is only as good as Qwen2.5... And I haven't tried Llama 405B because it's too expensive while not being optimized for coding.

1

0

0.000 POB

about 21 hours ago

RE: LeoThread 2025-02-21 13:43 Claude-sonnet but it's too expensive... Codestral 2501 is really good but not as good as Claude. DeepSeek-R1 is great when it's not doing infinite reasoning loops, and…

0

0

0.000 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 Speaking about AI Limitations, Reasoning models, at least DeepSeek sometimes go into Infinite Loop of reasoning: The user needs A but the user needs B but the user needs A but....

7

0

0.009 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 Is the answer to it known by humans?

3

0

0.001 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 Oooh... I think Hashtags are cool, though!!

3

0

0.001 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 I tried to summarize a video that only had Music and text, and the "summary" result was about the importance of Music!

3

0

0.001 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 Yeah, Twitter giving us Grok3 for Free is based move!!! (I imagine it'll be limited.)

3

0

0.001 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 I think open weights is as close as we'd get to Ope Source in the AI space.~

3

0

0.001 POB

1 day ago

RE: LeoThread 2025-02-21 13:43 Venice.ai's Premium image models are a lot better, @taskmaster4450le

3

0

0.001 POB