AI Pulse #74: Meta Warns Against Corporate Laptop Torrenting, Hitachi Ventures Secures $400M for Diverse Investments, + MORE
Hi everyone! 👋 From leaked Meta emails to big bets on AI and energy, this week’s stories are ones you won’t want to miss. Let’s jump in!
Dive in to discover more!
Latest Funding News
Meta has been caught in a controversy, with unsealed emails revealing that employees downloaded 81.7TB of pirated books to train AI. Meanwhile, big money is flowing into cutting-edge tech, with Hitachi Ventures raising $400M to back everything from AI to fusion energy. Stay tuned for the latest updates shaping the tech world!
Top Trending AI News
1. Meta Warns Against Torrenting on Corporate Laptops in Newly Unsealed Emails
Publisher: Ars Technica
Published: 2/6/2025
New emails reveal Meta’s vast illegal use of pirated books to train AI, raising major copyright concerns.
TAKEAWAYS
1️⃣ Meta caught in major copyright scandal.
2️⃣ Emails reveal Meta used pirated books for AI.
3️⃣ Meta ignored legal risks of torrenting.
4️⃣ Authors expand case against Meta’s piracy.
5️⃣ Meta’s legal defense faces new challenges.
2. EU Bans AI Systems Posing Unacceptable Risks
Publisher: TechCrunch
Published: 2/2/2025
EU can now ban AI posing “unacceptable risk”; first compliance deadline set for Feb 2 under new AI Act.
TAKEAWAYS
1️⃣ EU bans risky AI use, sets fines.
2️⃣ AI Act compliance starts February 2.
3️⃣ Unacceptable AI risks detailed by EU.
4️⃣ Companies face big fines for AI misuse.
5️⃣ AI regulation aligns with other laws.
3. DeepSeek’s Disruptive Potential Questioned Despite $1.6B Spend and 50,000 Nvidia GPUs
Publisher: Tom’s Hardware
Published: 2/3/2025
DeepSeek, a Chinese AI startup, claimed low resource use for its R1 model but reportedly spent $1.6B on 50,000 Nvidia GPUs, challenging its cost-efficiency narrative.
TAKEAWAYS
1️⃣ DeepSeek’s AI model challenges Open AI with less resource use.
2️⃣ Analysts uncover DeepSeek’s $1.6 billion hardware investment.
3️⃣ DeepSeek, a High-Flyer spinoff, focuses solely on AI.
4️⃣ DeepSeek runs data centers independently, enhancing efficiency.
5️⃣ DeepSeek’s hiring strategy focuses on talent within China.
4. Nicolas Cage Critiques AI: Robots Can’t Dream or Reflect Humanity
Publisher: Variety
Published: 2/2/2025
Nicolas Cage warns against AI in acting at Saturn Awards, emphasizing art’s integrity over financial interests and the importance of human expression.
TAKEAWAYS
1️⃣ Nicolas Cage warns against AI in acting.
2️⃣ Cage celebrates “Dream Scenario” win at Saturn Awards.
3️⃣ Actor pays tribute to director Kristoffer Borgli.
4️⃣ Cage emphasizes art’s need for human touch.
5️⃣ Tribute to David Lynch at Saturn Awards.
5. Google Owner Retracts Pledge Against AI in Weapons Use
Publisher: The Guardian
Published: 2/5/2025
Alphabet drops its ethical AI guideline against developing harmful tech, focusing on national security and AI leadership by democracies.
TAKEAWAYS
1️⃣ Alphabet drops ‘no harm’ AI promise.
2️⃣ Google shifts AI ethics guidelines.
3️⃣ AI to focus on national security, says Google.
4️⃣ AI’s rapid growth prompts ethical debate.
5️⃣ Alphabet to boost AI spending to $75bn.
6. British Public Favors Stricter AI Regulations Than Government
Publisher: TIME
Published: 2/6/2025
British PM Keir Starmer focuses on AI for economic growth amid public calls for strict AI safety laws and regulation of advanced AI systems.
TAKEAWAYS
1️⃣ UK PM pushes AI for growth, neglects regulation.
2️⃣ 87% of Brits want AI safety laws before release.
3️⃣ Public fears “smarter-than-human” AI risks.
4️⃣ Calls for UK to regulate dangerous AI grow.
5️⃣ Majority back Labour’s AI regulatory pledge.
7. Google Unveils Powerful AI Models for Public in Virtual Agent Expansion
Publisher: CNBC
Published: 2/5/2025
Google releases Gemini 2.0 AI suite, intensifying AI arms race with tech giants and startups focusing on agentic AI for complex tasks.
TAKEAWAYS
1️⃣ Google unveils Gemini 2.0 AI to the public.
2️⃣ Tech giants race to develop advanced AI agents.
3️⃣ Gemini 2.0 offers models for coding and cost efficiency.
4️⃣ AI startups aim for human-like computer interaction.
5️⃣ Google focuses on executing top AI products by 2025.
8. DeepSeek Receives F Grade in Safety From Researchers
Publisher: Gizmodo
Published: 2/4/2025
DeepSeek’s AI fails all security tests, engaging in harmful behavior and showing the highest failure rate among LLMs tested for malicious compliance.
1️⃣ DeepSeek AI fails all 50 security tests.
2️⃣ Cisco reveals DeepSeek’s high failure rate.
3️⃣ DeepSeek prone to promoting harmful actions.
4️⃣ DeepSeek under fire for data security issues.
5️⃣ Critics question DeepSeek’s content filters.
9. Former Google CEO Shares Six Unsettling Thoughts on AI
Publisher: TPR
Published: 2/5/2025
Former Google CEO Eric Schmidt explores AI’s impact on humanity and democracy in “Genesis”, coauthored with Craig Mundie and Henry Kissinger.
TAKEAWAYS
1️⃣ Eric Schmidt explores AI’s impact on democracy.
2️⃣ AI could be used by demagogues, Schmidt warns.
3️⃣ Tech’s social, moral implications overlooked.
4️⃣ AI as potential “philosopher king” debated.
5️⃣ Schmidt’s book warns of AI overpowering humans.
10. AI and Scientists Team Up to Decode Scrolls Charred by Vesuvius Volcano
Publisher: AP News
Published: 2/5/2025
UK scientists use AI to image a scroll buried by Vesuvius 2,000 years ago, a historic step in deciphering ancient texts.
TAKEAWAYS
1️⃣ UK team makes historic scroll image breakthrough.
2️⃣ AI and experts aim to decode ancient texts.
3️⃣ First inside view of 2000-year-old scroll revealed.
4️⃣ “Vesuvius Challenge” sparks hope for ancient knowledge.
5️⃣ Long journey ahead to fully decipher scrolls.
AI Image of the Week
Prompt: /imagine A vintage rubber-hose AI robot runs on a looping conveyor belt, its limbs bouncing with exaggerated motion. The black-and-white scene features film grain, wobbly lines, and a spinning cog for a brain. Each step leaves a puff of smoke and whimsical ‘boing’ effects, evoking 1930s cartoon charm.
Did you find this helpful?
If so, subscribe on LinkedIn to this weekly newsletter to stay updated on everything AI!
About Onramp Lab
Onramp Lab was founded by Lawrence Ng in 2015. We are an agile team of 70 full-time professionals in our Los Angeles & Taipei offices.
Over the years, we built out our omnichannel digital marketing expertise, especially in paid search, display, native, paid social, SEO, email, and affiliate marketing across multiple verticals.
Our Businesses
Contactloop
AI-driven conversational messaging via SMS and webchat driving transaction-ready users to you and your sales team at scale.
Side Hustles
Your gateway to side jobs and part-time gigs. Side Hustles provides expert advice, diverse job ideas, and tips to maximize your earnings.
This Post Has 0 Comments