@matthewberman_ai: I asked Kimi K3.1 Pro to create a graph showing its benchmarks. It returned a chart with scores for coding, reasoning, mathematics, and agentic tasks. The first thing I noticed was how broad those categories were. They didn't tell me which benchmarks the numbers came from, and the scores didn't look accurate. So I asked whether the numbers were actually accurate. The response was: "No. They are invented placeholders." #AI #Future #LLM #Kimi

matthewberman
matthewberman
Open In TikTok:
Region: US
Saturday 24 October 2026 23:42:24 GMT
1961
52
5
0

Music

Download

Comments

calvinhlungwani
KC Steeper :
do the same with opus5.5 and GPT and see the bias
2026-10-25 14:59:02
0
user6385528258191
AD :
😁 whoa the age of AI
2026-10-25 10:58:51
0
meganastyrich
Cat in the Hat :
Welcome to AI. Most days it is junk
2026-10-24 23:58:40
0
digital1income1
Toqeer 🇨🇦 :
I think more people need to understand what financial independence actually means, and @Steve J does a great job bringing attention to the subject. Definitely worth checking out if FIRE is something you’ve been researching.
2026-10-26 02:03:44
0
To see more videos from user @matthewberman_ai, please go to the Tikwm homepage.

Other Videos


About