Harish

10 posts

Harish

Harish

@HarishM687217

Katılım Mayıs 2026
38 Takip Edilen1 Takipçiler
Harish retweetledi
Umesh Kumar Yadav
Umesh Kumar Yadav@Umesh__digital·
Each time claude releases a new model…
English
52
16
499
30.7K
Harish retweetledi
Thomas Ricouard
Thomas Ricouard@Dimillian·
Here is a walkable demo of the scene in Blender. Without lighting, but this is still amazing what 5.6 Sol did in two passes. It took less than 45 minutes on Ultra.
Thomas Ricouard@Dimillian

I've asked 5.6 Sol to push a bit the geometry accuracy and materials details, and it's surprising me. This is a scene that 5.6 all built in Blender and that can be manipulated and walked around. And it keeps generating tastful static screens out of it.

English
15
15
226
42.1K
Palak🎀
Palak🎀@Palak3312·
Be honest:- How many programming languages can you actually code in without AI? 👀
Palak🎀 tweet media
English
56
3
73
3.5K
komal sharma
komal sharma@itz_komal01·
This is 100% true...
komal sharma tweet media
English
9
2
48
3.3K
Harish
Harish@HarishM687217·
Harish tweet media
ZXX
0
0
0
8
Vaibhavii
Vaibhavii@Vaiibhavii_31X·
Average vibe coding session 👀
Vaibhavii tweet media
English
12
1
102
6.1K
Harish
Harish@HarishM687217·
If humanity had refused to adopt new technologies in the past, many of the technologies we rely on today would never have existed. Progress happens because each generation builds upon the innovations of the previous one.
Harish tweet media
English
0
0
0
15
Elon Musk
Elon Musk@elonmusk·
Try Grok Build, it’s awesome! X.ai/cli
X Freeze@XFreeze

Grok 4.5 now also ranked #1 on the Long-Horizon Terminal-Bench by binary pass rate, outperforming Claude Fable 5, Claude Opus 4.8 and GPT-5.6-sol Under the strictest scoring metric - where a task counts only if it is fully solved with a perfect reward and zero errors......Grok 4.5 finished clearly ahead of every other tested model This matters for real-world coding, automation and difficult engineering work because long-horizon terminal tasks require much more than making partial progress The model has to maintain context, recover from mistakes and successfully complete an entire workflow across hundreds of steps Grok 4.5 is showing serious strength on complex agentic tasks over extended periods of time

English
2.5K
2.3K
10.1K
7.6M