Xiao Lu @xiaolu_sprmind - Twitter Profile

16 minutes ago

Mistral just open-sourced a ~1T-parameter model trained in Europe. Not #1 on leaderboards, and that's fine: for banks and governments the question is which strong model you can own and run in-house. #AI #SovereignAI https://t.co/YGuRmRRYfG

xiaolu_sprmind's tweet photo. Mistral just open-sourced a ~1T-parameter model trained in Europe. Not #1 on leaderboards, and that's fine: for banks and governments the question is which strong model you can own and run in-house. #AI #SovereignAI
https://t.co/YGuRmRRYfG https://t.co/aVSDkscX9T

0

5

Xiao Lu

@xiaolu_sprmind

about 16 hours ago

Google's top-ranked model, Gemini 4 Carbon, goes to cyber defenders first, with wider access phased in later. Right call for a dual-use model. The real test: how fast protection reaches the long tail of smaller orgs. #AI #Cybersecurity

xiaolu_sprmind's tweet photo. Google's top-ranked model, Gemini 4 Carbon, goes to cyber defenders first, with wider access phased in later.

Right call for a dual-use model. The real test: how fast protection reaches the long tail of smaller orgs.

#AI #Cybersecurity https://t.co/jZRbTGgxwK

1

0

20

Xiao Lu

@xiaolu_sprmind

about 20 hours ago

Similar AI spend, opposite cash result. Microsoft: $45.2B capex, $18.6B free cash flow. Alphabet: $51.4B capex, -$9.1B free cash flow. The market's AI test has changed: not "are you spending enough?" but "is the spending funding itself yet?" #AI #Finance

xiaolu_sprmind's tweet photo. Similar AI spend, opposite cash result.

Microsoft: $45.2B capex, $18.6B free cash flow.
Alphabet: $51.4B capex, -$9.1B free cash flow.

The market's AI test has changed: not "are you spending enough?" but "is the spending funding itself yet?"

#AI #Finance https://t.co/OBdB9CWQUW

1

0

1

14

Xiao Lu

@xiaolu_sprmind

1 day ago

More than 40% of GitHub pull requests now start with an AI agent's first draft, up from 17% in March. The bottleneck is moving from writing code to review, memory and ownership. #AI #AIAgents

xiaolu_sprmind's tweet photo. More than 40% of GitHub pull requests now start with an AI agent's first draft, up from 17% in March.

The bottleneck is moving from writing code to review, memory and ownership.

#AI #AIAgents https://t.co/3MncrWjF0U

0

8

Xiao Lu

@xiaolu_sprmind

1 day ago

S&P 500 Q3: index EPS growth expected around 27%. The median company: about 9% (Goldman Sachs estimates). Top 10 contributors drive 68% of the growth. The headline tells you how AI infrastructure is doing, not the whole economy. #Earnings #AI

xiaolu_sprmind's tweet photo. S&P 500 Q3: index EPS growth expected around 27%. The median company: about 9% (Goldman Sachs estimates). Top 10 contributors drive 68% of the growth.

The headline tells you how AI infrastructure is doing, not the whole economy.

#Earnings #AI https://t.co/zClv4gMOMY

0

6

Xiao Lu

@xiaolu_sprmind

2 days ago

The Netherlands plans about 25,000 fewer central government jobs by 2035, with AI part of the efficiency drive. The AI plan itself is due in Q1 2027. Target first, plan second. In public services, an AI mistake lands on a citizen, not a cost line. #AI #AIGovernance

xiaolu_sprmind's tweet photo. The Netherlands plans about 25,000 fewer central government jobs by 2035, with AI part of the efficiency drive. The AI plan itself is due in Q1 2027.

Target first, plan second. In public services, an AI mistake lands on a citizen, not a cost line.

#AI #AIGovernance https://t.co/Ja7opEsOTj

0

9

Xiao Lu

@xiaolu_sprmind

2 days ago

Nvidia is putting $3B into Meta's new equity offering. AMD adds $1B. Meta is raising it to build AI infrastructure, which means buying a lot of chips. Not a scandal. But it makes the AI demand signal noisier: the shovel sellers now own part of the mine. #AI #Semiconductors

xiaolu_sprmind's tweet photo. Nvidia is putting $3B into Meta's new equity offering. AMD adds $1B. Meta is raising it to build AI infrastructure, which means buying a lot of chips.

Not a scandal. But it makes the AI demand signal noisier: the shovel sellers now own part of the mine.

#AI #Semiconductors https://t.co/pM05QiiF3o

0

27

Xiao Lu

@xiaolu_sprmind

6 months ago

https://t.co/SzG7BXauje Hermes Agent: Building Persistent, Self-Improving AI Systems is a practical guide to one of the most important shifts in modern AI: the move from session-based assistants to persistent agents that learn, adapt, and improve over time.

1

0

81

Xiao Lu

@xiaolu_sprmind

8 months ago

Mastering OpenClaw is a practical, engineering-first guide to building privacy-first local AI agents—systems that run on your own machine, keep sensitive data under your control, and still feel fast, capable, and production-ready. https://t.co/iVMymF6f0J

0

44