_Singularity001

82 posts

_Singularity001 banner
_Singularity001

_Singularity001

@_Singularity001

Digital content/ AI prompts/ AI creator Solving 90% of my problems with AI. Procrastinating on the other 10%. I’m not smart, I just use AI.

37°14′0″N 115°48′30″W 加入时间 Kasım 2025
55 关注11 粉丝
_Singularity001
_Singularity001@_Singularity001·
I hope @xai team add Grok 4.3 to the API soon🙏
English
0
0
0
18
LudovicCreator
LudovicCreator@LudovicCreator·
Share your favorite generation NOT POSTED this week Any tool , image or video ... I will quote and pin Monday the one with more likes
LudovicCreator tweet media
English
81
11
147
3.5K
Mario Nawfal
Mario Nawfal@MarioNawfal·
It's happening, folks. That dystopian Netflix drama about AI girlfriends is only months away from becoming a reality. If you have to get one, make sure yours is powered by Grok.
English
468
191
1.6K
475.7K
_Singularity001
_Singularity001@_Singularity001·
She did the Robot doll 3.0 She really did an amazing job!!
English
0
0
0
29
_Singularity001
_Singularity001@_Singularity001·
@techdevnotes I have no doubt that the xAI team will ship something that is not just good but great!
English
0
0
0
39
Tech Dev Notes
Tech Dev Notes@techdevnotes·
What if Grok Build turns out to be surprisingly good
English
52
5
244
10.2K
Max ✦
Max ✦@MaxVOAO·
Give me your best advice
Max ✦ tweet media
English
12
4
45
2.2K
_Singularity001
_Singularity001@_Singularity001·
Google Is Rebuilding AI Compute. 👀 What’s unfolding: Google isn’t simply releasing TPU 8, it’s re-architecting its entire AI stack. From silicon to networking to storage, everything is being redesigned. By separating hardware for training and inference, and integrating its own Arm-based CPUs, Google is pulling more of the pipeline in-house and reducing dependence on outside vendors. 🌍 Why it matters in the real world: Right now, most AI workloads lean heavily on Nvidia GPUs, which drives up costs, tightens supply, and limits how much cloud providers can differentiate. Google’s approach is an attempt to break out of that constraint, owning the full stack could mean lower costs, better efficiency, and more control over scaling models and serving massive demand. 🛎️ The bottom line: This is about eliminating Nvidia as a choke point and turning compute into a strategic advantage. If it works, Google doesn’t just save money, it gains pricing leverage, tighter margins, and a more defensible position in the race to power large-scale AI.
_Singularity001 tweet media
English
0
0
1
18
_Singularity001
_Singularity001@_Singularity001·
I am a man of simple tastes, I'm always satisfied with the best. -Oscar Wilde Grok Imagine
English
0
1
2
88
_Singularity001
_Singularity001@_Singularity001·
The Big Pivot: DeepSeek Goes Hunting for Capital DeepSeek is reportedly breaking its streak of independence, seeking at least $300 million in external funding at a valuation north of $10 billion. This is a massive strategic U-turn for a company that previously branded itself as a scrappy, investor-resistant outsider. The move isn't a coincidence. With V4 facing multiple delays and rumors swirling that rivals like OpenAI and Anthropic are successfully "gatekeeping" the high-quality data DeepSeek needs for distillation, the era of cheap shortcuts appears to be ending. DeepSeek is no longer just a lean research lab; it’s a company realizing that original scale requires original capital. The "Distillation" Wall: DeepSeek’s rapid ascent was largely fueled by distillation—a process that is significantly faster and more affordable than training a model from scratch. However, this method has a built-in shelf life: Dependency: Distillation only works if a "frontier" company spends billions to build the initial model. Access Restricted: As competitors tighten API rate limits and introduce legal safeguards, the flow of "imitation data" is drying up. Originality Gap: You cannot indefinitely climb the AI ladder by simply mimicking the reasoning traces of others. Eventually, every serious contender has to pay the "innovation tax"—investing in their own massive compute clusters, top-tier researchers, and proprietary breakthroughs. 🛎️ The Bottom Line: DeepSeek has officially hit the limits of imitation. The upcoming V4 release is no longer just a benchmark test; it is a proof of concept for the company’s survival. DeepSeek must now demonstrate it can build a frontier-grade model using its own infrastructure, unique ideas, and hardware stack. The Verdict: DeepSeek is at a crossroads. It must evolve into a primary innovator or risk being remembered as the world’s most efficient follower in a race that only honors the leaders.
English
0
0
0
73
_Singularity001
_Singularity001@_Singularity001·
@cb_doge I don’t see the option to do that? And my app is up to date.
English
0
0
7
209
DogeDesigner
DogeDesigner@cb_doge·
X just launched Custom Timelines — pin any topic to your home feed and get a fully personalized feed powered by Grok. Here's how to set it up:
English
98
162
655
51.1K
Tech Dev Notes
Tech Dev Notes@techdevnotes·
Read Aloud feature in Grok Web now has an Option to choose a Different Voice
Tech Dev Notes tweet media
English
3
2
76
3.9K
Tech Dev Notes
Tech Dev Notes@techdevnotes·
SpaceX / xAI and Cursor are now Working Closely
SpaceX@SpaceX

SpaceXAI and @cursor_ai are now working closely together to create the world’s best coding and knowledge work AI. The combination of Cursor’s leading product and distribution to expert software engineers with SpaceX’s million H100 equivalent Colossus training supercomputer will allow us to build the world’s most useful models. Cursor has also given SpaceX the right to acquire Cursor later this year for $60 billion or pay $10 billion for our work together.

English
9
8
192
8.4K
_Singularity001
_Singularity001@_Singularity001·
The mothership soars above a metropolis city at twilight. @grok Imagine
English
1
0
2
105
_Singularity001
_Singularity001@_Singularity001·
There are ancient portals on this Earth, still open for our transformation and learning.
_Singularity001 tweet media
English
0
0
0
16