They do this EVERY time. Why do people keep giving them money,
2026-11-05 10:20:07
3
Jorge Santos :
It's getting insanely bad!!
2026-11-04 20:44:38
25
:
That’s not true at all
2026-11-05 12:02:47
0
Pascal :
its still 10/10 , perfect work 🤷
2026-11-04 22:27:10
7
colin.tm :
local LLM
2026-11-04 22:48:28
0
Linux :
It still Perfect, I Test GPT 6.1 and it was terrible in agent task, just switched too Opus back and it was Perfect
2026-11-05 01:46:20
1
Gustas :
I just keep using antigravity with flash 3.8 it ain't the best but it doesn't have many issues
2026-11-04 22:55:24
1
𓆩 ʍ e w 𓆪 :
when calls to models increase they get quantized thats all it is, no big scams just the fact they need to quantize them to keep them up
2026-11-04 23:25:16
0
Tetris_Prime :
I know this is a very disliked opinion, but building your workflow on local models like Gemma4 26b on a 2090 will provide you with much more continuity and much less risk.
2026-11-04 15:50:56
3
javis :
astra is so goated
2026-11-04 16:14:51
0
Matti :
It’s still fantastic
2026-11-05 01:04:23
1
cpy :
Had that issue with sol recently. Now its ok again
2026-11-05 07:17:00
0
rnyrgs :
Last week in low effort disappeared so fast, and last 3% disappeared when even no sessions were active :(
2026-11-05 07:01:32
0
David_hrzn :
I bet there will be a new Model before the ipo
2026-11-05 07:40:32
0
YourFather :
i notice that on gpt
2026-11-05 07:11:14
0
jonhionis :
So many mistakes man it feels like opus 5...
2026-11-05 08:50:48
0
MigBroo :
Grok
2026-11-05 06:06:20
0
ffugyctxxyvvvy :
2026-11-05 08:17:18
0
𒉭ⒺⓇⒾⓀᣞ :
yeahhh its doing really shety
2026-11-04 20:01:17
1
gianmt :
Enshittification it’s everywhere my friend
2026-11-04 12:54:54
2
Steve Woodwill :
I agree it's thick as pigshit now
2026-11-05 02:14:12
1
Confused :
U cant just say that. Do you have any proof for that? Like actual data?
2026-11-05 08:16:26
0
△ ᛋᛅᚱᛒᛅᚾᛏᛋ ᛒᛁᛏᛅ ▽ :
canceled pro abo ^^ dont need that anymore and make more myself again
2026-11-05 09:55:03
0
mhtroylog :
@albert_olgaard
2026-11-05 12:33:04
0
To see more videos from user @aiwithzygis, please go to the Tikwm
homepage.