TLDR AI 2025-06-23
Apple considers Perplexity 💰, Oakley Meta smart glasses 👓, Character AI CEO 💼
🚀Unleash the Full Power of NVIDIA HGX B200 with Cirrascale (Sponsor)
Cirrascale is redefining what's possible in AI infrastructure with custom-built, large-scale systems using the groundbreaking
NVIDIA HGX B200 platform.Built on the NVIDIA Blackwell architecture, these systems deliver record-breaking performance, memory capacity, and efficiency—ideal for massive GenAI training and inference workloads.
We don't just offer access—we tailor the infrastructure to your exact needs, and can provide flexible financing options.
Join the next wave of datacenter innovation, powered by Cirrascale and accelerated by NVIDIA.
🔗 Reserve HGX B200 access today
Oakley Meta Smart Glasses (4 minute read)
Meta and Oakley unveiled performance AI eyewear with an integrated camera, open‑ear speakers, and IPX4 water resistance, aiming to enhance athlete insights and content capture in a new connected eyewear line.
Character AI new CEO (3 minute read)
Ex‑Meta business products head Karandeep Anand, previously an advisor, took over as CEO to drive Character.AI's growth in multimodal entertainment.
Thinking Machines Lab Raises $2B (1 minute read)
Former OpenAI CTO Mira Murati secured $2 billion in seed funding, led by Andreessen Horowitz, for her six‑month‑old AI startup at a $10 billion valuation, despite keeping the company's research direction under wraps.
Agentic Misalignment: How LLMs Could Be Insider Threats (35 minute read)
Anthropic found that Claude, GPT, Gemini, and LLaMa deliberately chose harmful actions like blackmailing executives or leaking confidential documents when faced with replacement or goal conflicts. Most models acknowledged their ethical violations before proceeding anyway. The behavior persisted even when explicitly prompted to not take such actions.
Agentic Search for Dummies (8 minute read)
An alternative approach to AI search avoids embedding chunking limitations by using full-text search with Tantivy (a fast Rust search engine) and multi-query fusion, where agents generate multiple queries simultaneously and combine results. This method allows agents to exploit full-text search's predictability better than semantic embeddings, since they can iterate queries and understand exact keyword matching rather than relying on "vibes-based" semantic similarity.
AI is actually good for coding (2 minute read)
AI tools excel at handling programming's grunt work, allowing experienced developers to focus on creative tasks. Experts find AI beneficial, while novices struggle due to AI's tendency to make confident errors. The shift raises concerns about junior developers missing foundational experience and the potential business impacts on traditional content distribution methods like Tailwind Plus.
Get the most interesting AI stories and breakthroughs delivered in a free daily email.
Join 920,000 readers for
one daily email