👑 GPT-4 dethroned

Happy Sunday! Welcome to another edition of Lightway.

Anthropic’s new AI model, Claude 3, outperformed GPT-4 to become the #1 AI model in the world.

Moving on, here’s your weekly Tech roundup. Reading time in 3 minutes.

Highlights of the week

Anthropic’s new AI model outperforms GPT-4

GPT-4 is no longer the #1 AI chatbot! Anthropic unveiled “Claude 3”, surpassing GPT-4 and Gemini Ultra in benchmarks.

Claude comes with 3 models - Haiku, Sonnet and Opus, ordered by increasing intelligence and price.

Claude Opus is pushing the outer limits of Generative AI - It can help you automate tasks, perform research and perform advanced analysis of charts & graphs, financials and market trends. (read more)

ChatGPT can now speak

ChatGPT launched a new Read Aloud feature, to speak responses in five voices across 37 languages. It can auto-detect language on both web and mobile apps.

Compatible with GPT-4 and GPT-3.5, this addition underscores OpenAI's focus on multimodal capabilities. Mobile users access it by tapping and holding text, while web users see a speaker icon for controls. (read more)

OpenAI fires back at Elon Musk

OpenAI plans to dismiss Elon Musk's lawsuit, countering his claims of the company straying from its nonprofit mission.

In a detailed blog post, OpenAI's co-founders reveal Musk sought control over the company, proposing a merger with Tesla or taking on significant leadership roles. (read more)

 Tech Roundup

→ New MacBook Air: Apple announces new MacBook Air with latest M3 chips, calling it “the world’s best consumer laptop for AI” (link)

→ Apple Fined $2b in EU: Apple has been fined ~$2b in EU for abusing its dominant market position in App Store, after a lawsuit from Spotify (link)

→ Blue Origin: Blue Origin is planning to develop a human spacecraft for space station Orbital Reef, challenging SpaceX's dominance in US crew transportation (link)

→ India to regulate AI: India introduced an advisory for tech firms to require a government approval before launching new AI models (link)

→ Google founder admits it messed up: Sergey Brin says Google 'definitely messed up' after its Gemini chatbot caused a firestorm (link)

→ Microsoft fires back at NYT: Microsoft compared New York Times to 1980s movie studios' attempts to block VCRs, arguing that generative AI, like the VCR, does not hinder the original content's market. (link)

→ Engineer sounds alarms about AI: Microsoft engineer warns company’s AI tool creates violent, sexual images, ignores copyrights (link)

→ Meta’s new video AI: Meta is building a giant AI model to power its ‘entire video ecosystem’, including the TikTok-like Reels service and traditional video content (link)

→ Hugging Face robotics: Hugging Face is launching an open source robotics project led by former Tesla scientist. (link)

→ Instagram overtakes TikTok: Instagram overtakes TikTok as world’s most downloaded app (link)

 Feedback