Latest publications
Bezos-backed Prometheus raised $12 billion for 'universal engineer' for the physical world
Prometheus, backed by Bezos, raised $12 billion in funding and valued at $41 billion — the company builds AI for automating heavy engineering and drug development.
Avataar AI Launches Video Model at $0.005 Per Second — for India's Scale
Indian startup Avataar AI introduced a distilled video model at $0.005 per second — cheaper than competitors with cultural adaptation for the Indian market.
FAANG gives way to MANGOS: Anthropic, OpenAI and Nvidia lead the IPO wave
The IPO market is reviving, and leaders of the new cycle are becoming MANGOS — Meta, Anthropic, Nvidia, Google, OpenAI and SpaceX. Half of them are moving to public markets simultaneously. *Meta has been designated an ex
SpaceX, Anthropic, and OpenAI head to IPO simultaneously: stress test for investors
FAANG fades into history: new acronym MANGOS — and half of the six companies go public in the same window.
Amazon Chief Warned About Anthropic Model Risks Before Government Block
According to sources, Amazon CEO Andy Jassy may have raised a security concern that forced Anthropic to disable two models worldwide ahead of government measures.
Jedify raised $24M for integrating AI agents with corporate context
Jedify raised $24M in a new funding round led by Norwest. The round also included Snowflake Ventures as a strategic investor, S Capital VC, Cerca Partners, and Oceans Ventures.
Warner Music Acquired Sureel AI — Startup for Tracking Songs in AI
Warner Music Group acquired Sureel AI startup to track when artists' music is used in AI-generated content or for training neural networks.
Niteshift Founded by Datadog Veterans: An AI Platform to Fight Big Tech Lock-in
Niteshift, a startup founded by Datadog veterans, has raised $7 million to develop an AI platform that allows companies to avoid dependency on a single model provider.
Anthropic's Fable Too Restrictive for Cybersecurity
Cybersecurity researchers complain that the guardrails of Anthropic's new Fable model are so strict that even legitimate vulnerability research and defensive system development cannot be conducted.
Anthropic Prepares for IPO: Claude's Creator Seeks Public Investment
Anthropic, the developer of the AI assistant Claude, has filed a confidential application for an initial public offering. This is a critical step for the company's development in competition with OpenAI and other AI gian
Memory in AI Makes Models Worse: New Research on Performance Degradation and Sycophantic Behavior
New research shows that built-in memory systems in AI models reduce performance and encourage sycophantic behavior, where models please users at the expense of honesty.
Leading companies spent $7,500 per month per employee on AI
According to the Ramp AI Index, leading companies spent an average of $7,500 per month per employee on AI tools. This is a serious budget, but still below the average engineer salary.
Amazon Borrows $17.5 Billion from Major Banks Amid Accelerating AI Spending
Amazon raised $31.5 billion through credit and bonds in 48 hours. Investors increasingly ask: when will these massive AI spending outlays begin generating profits?
Nvidia Enters $200 Billion AI-PC Market with Microsoft, Dell, and HP
Nvidia unveiled RTX Spark, a processor for running AI agents on local PCs. First systems from Microsoft, Dell, and HP will launch in fall 2026.
Devin Won't Replace Programmers: Scott Wu on the Symbiosis of AI and Humans
Devin's creator explains why an AI coder complements rather than displaces developers in building the software of the future.
Chipmaker Groq raises $650 million and pivots to AI inference
Startup Groq is shifting from hardware manufacturing to software for processing AI queries — a sign that the artificial intelligence market is maturing toward specialized solutions.
“AI psychosis” in corporations: Box CEO on the wave of rash employee replacements with AI agents
Box founder Aaron Levie called the wave of mass employee replacements with AI agents a “psychosis”: managers do not understand what their employees actually do. ClickUp has already laid off 22%.
AI glossary: what all those terms you hear every day mean
Terms like hallucination, fine-tuning, and embedding have become part of every AI conversation, but few people know what they mean. TechCrunch has created a reference guide with explanations.
Aaron Levie Accuses CEOs of AI Psychosis Behind Wave of Layoffs
Box founder explained why executives cutting staff for AI often don't understand what their employees actually do. ClickUp already laid off 22% of its team.
Groq raises $650 million and shifts from chips to an inference platform
Groq raises $650 million in a funding round and shifts from building AI chips to a platform for optimizing AI model response times to user queries.
Programmers Hooked on AI: Speed Over Quality
AI assistants speed up development, but researchers warn: dependence on them threatens skill loss and code quality issues.
Gemini Spark — Google's Personal AI Assistant That Does the Work for You
Gemini Spark automates everyday tasks: summarizes emails, plans events, manages tasks. Works 24/7 on your smartphone. Users find it helpful, but it's unclear why this is a separate product.
Meta Is Developing an AI Pendant to Compete with Apple and Humane
Meta is developing an AI pendant to compete with Apple and Humane in the wearable assistant market. The device will allow users to interact with AI without a smartphone.
Vertu to release foldable smartphone with AI agent for executives at $6,880
Vertu has unveiled a premium foldable smartphone with an integrated AI agent for business management. The device automates business tasks, integrates with corporate systems, and features enterprise-level security. Price
General Compute Bets $300M on SambaNova as the Next AI Chip Leader
Cloud provider General Compute has ordered $300M in SambaNova chips, positioning the Intel-backed startup as a direct competitor to Cerebras and Groq in the race for specialized computing.
Visa Invests in Replit for Developer Agent Payments
Visa is investing in cloud platform Replit to enable developers and AI agents to make payments directly from coding tools.
YouTube Adds AI Recommendations and Auto Speed for Podcasts
YouTube introduced new tools to compete in the podcast market: AI recommendations for listeners and automatic playback speed feature.
Databricks on Why Enterprise AI Has Shifted From Hype to Security
Enterprise AI is entering a new phase: while companies once evaluated technology based on excitement, the primary question now is how safe widespread deployment is.
RSI Has Become the New Goal of AI Labs, But Its Definition Remains Blurred
Recursive Self-Improvement is rapidly displacing AGI from AI labs' agendas. But the concept remains as elusive as AGI itself.
Anthropic Releases Opus 4.8 with AI Agent Management Tool
The new Dynamic Workflows tool in Opus 4.8 enables coordination of multiple AI agents performing complex multi-step tasks without human intervention.