Ty’s AI’s
Region: US
Wednesday 03 June 2026 05:13:21 GMT
29857
1526
105
113
Music
Download
Comments
melinzmazur :
Pair Blackwell 6000 🥰
2026-06-04 21:54:37
2
foodfighter08 :
yeah just spend $20,000 on hardware
2026-06-03 16:34:17
13
CobyTheRealtor :
clearly you know ball
2026-06-26 21:30:27
2
Michael :
how does Hermes compare on pricing? I’ve been using Workbeaver and it’s been pretty good for the amount of automation I get out of it
2026-06-18 13:43:46
52
f744uasb219 :
based. thoughts on rowboat?
2026-06-03 14:04:43
0
Ty’s AI’s :
Add firecrawler and any other GitHub repo dedicated for agents and its literal agentic gold
2026-06-03 05:29:30
23
zioshill :
The only local LM model that worked reliably for me are open whisper for speech and image analytics with YOLO V11. an open wake word for a fast wake word detection. for llm, qwen3.6:27b on an rtx 3090
2026-06-05 17:06:09
1
Victor Caballero :
I just got my rtx 5090 and 128 gb to run qwen and nemotron. Will be running locally ! I’m gonna have to look into this Hermes
2026-06-07 15:07:50
6
Mikael Söderberg :
I think you are right. An alternative is also DeepSeek + Hermes.
2026-06-17 17:30:57
2
hepek :
what size Qwen3.6 ?
2026-06-10 21:54:28
1
Thias :
I've been using Ollama Gemma4 31b-cloud for most non work stuff
2026-06-05 12:40:37
2
Alex :
GLM 5.2 + Hermes + Hancho*
2026-06-27 16:32:38
1
Bo Bell :
I just bought a new MacBook Pro for this 48gb ram m5 chip but not confident what’s the best local model to run
2026-06-05 23:57:42
0
jeffs2incher :
Would we be able to get to a point where models would be able to run on normal laptops? Is that ever a possibility cause having to buy machines that are beefy in resources is a bit strenuous
2026-06-04 03:40:26
0
Secret Rap Battle :
How has all innovation stopped and we’re stuck right here when it just advanced between now and six months ago and I’m doing stuff right now I couldn’t do 180 days ago
2026-06-06 06:12:22
1
Avery :
When AI slows down it will make sense to invest in local hardware.
2026-06-05 22:01:22
0
user9528471202904 :
yep I'm getting 27 tokens a second on a $1500 laptop
2026-06-14 01:13:29
1
kellenchase :
I’ve been rocking kimi k2.6. Should I check out Qwen?
2026-06-07 06:07:05
1
MetaMoney :
Ram or vram?
2026-06-08 18:21:48
0
rickeyhendersons :
The spark laptop will do that combo its rightful justice
2026-06-06 22:41:41
1
Otis Carlisle :
It’s definitely not even close to stopping. But yeah I agree. I snagged a 5090 back when It 2k. I use that with open code. It’s as good as any model we had this time last year.
2026-06-19 19:43:59
1
Tanner Booker :
3.6 models naw they are like small routing agents for me i run on my cpu for my bigger 120b models and stuff. maybe for some people but for like real work models like that are way to small to really do anything meaningful with
2026-06-04 08:52:53
3
TiX :
been running this since 3.6 came out. it's fantastic
2026-06-21 03:09:18
1
quovadisveritas :
Most coding can be done with an excellent harness
2026-06-10 00:28:28
2
Sandrinho4x4 :
"Why rent a house if tents exist?"
2026-06-03 15:34:49
2
To see more videos from user @tys.ais, please go to the Tikwm
homepage.
