Ty’s AI’s
Region: US
Wednesday 10 June 2026 19:28:33 GMT
61861
2733
134
151
Music
Download
Comments
Terv338 :
Dont use ollama. use lama.cpp its more efficent
2026-06-11 22:52:22
96
sionlockett :
Gemma 4 is almost unusable in coding tasks locally for me, or maybe I'm just spoilt using gpt 5.5
2026-06-11 11:30:08
51
ShiCkM :
I think ollama 3.1 wasnt great. any other odeas😁
2026-06-11 21:29:22
0
Matthew Thorsen :
Interesting, I’ve found phi4:14b on my 12gb vram to be superior and faster. Give it a test if you haven’t yet. Also running through ollama
2026-06-27 16:39:12
1
. :
You can run qwen 3.6 35b MoE if you optimize it I am running it at 33 tokens with some optimizations on a rtx 4060 And for me gemma 4 vs qwen is like gemini vs Claude qwen is smarter but gemma is more creative
2026-06-13 06:00:56
11
Not So Friendly Local AI Dev :
i wouldn't say its the craziest. qwen 3.6 27b is the craziest. gemma 4 12b is cool tho
2026-06-14 04:31:18
3
epalacaswebmaster :
Is it good for vibe coding?
2026-06-12 23:15:22
0
AI For People Tired of Noise :
Running it on ollama? Great until you run out of usage unless you pay or go full local. 🔥
2026-06-14 00:23:42
0
mirth_is_king :
im trying to see what llms will work on an old dell r530 enterprise server.
2026-06-12 19:37:36
0
Ricardo Gonzalez :
bro laptop on the carpet is wild 💀
2026-06-13 22:29:43
3
I'm SbHaAnDnOeWd :
it's effortless to set it up. why did you need ai for it?
2026-06-15 20:16:09
1
houariZEN :
i did use a lot of local ia they are just no match to llm company i switched to deesek v4 flash wayyy more better for cheap
2026-06-11 17:23:11
0
claudecto :
This will only be truly useable when an 8gb laptop can run a 12b uncensored model on Hermes with actually useable tool calling
2026-06-14 20:10:57
1
Solaitions.com :
can u admit it doesn't do the job the 12b model
2026-06-11 07:16:40
0
:
is this better than fable 5? and can i use it to code whatever without safeguards?
2026-06-15 03:22:58
0
Irakli Dzadzamia :
Could you please give me some tips on common beginner mistakes and things that are good to know from the start?
2026-06-22 20:31:26
0
ToddMichaelTV :
I can't get anything done agenticlaly with local models. they just want to chat and tell me what to do ✌️😅
2026-06-17 01:09:30
0
Gluteus Maximus :
What do you think about Odysseus?
2026-06-15 19:46:43
0
babulono :
Я подключил его к open code на больших задачах у меня просто прерывается запрос, результат не возвращает.
2026-06-13 08:26:40
0
Gladiator blud🗿🥹 :
qwen 3.5 is better
2026-06-11 17:53:04
4
username0010110020201 :
its not
2026-06-11 07:03:57
0
🤷♂️ :
What do you use AI for? Like what is it programming the whole time?
2026-06-10 19:57:53
1
chris4128793 :
To do what?
2026-06-12 23:32:44
1
🇨🇦 Alan Pintor :
I just tested it on my openclaw and it is the best so far, thanks 🤝
2026-06-14 05:57:58
1
minuterepeater :
I used gemma and it just kept replying with "Hungary"
2026-06-11 07:21:48
1
To see more videos from user @tys.ais, please go to the Tikwm
homepage.
