GPT-5 router backlash — OpenAI reverts “One Model” auto-routing after user complaints, restoring Fast, Thinking, and GPT-4o modes.

Google AI search update — Users can now set “top sources,” shifting traffic dynamics as AI summaries reduce clicks.

Claude 4 Sonnet expands — Anthropic boosts developer appeal with 1M-token context window for massive API workloads.

Local AI hosting — OpenAI’s GPT-OSS 20B now runs offline on Windows with GPU acceleration via Ollama/LM Studio.

Perplexity’s bold move — Startup bids $34B for Chrome, pledging $3B to Chromium and open-source compliance.

Meta AI’s evolution — Model shows self-improving behaviours, fuelling debate on paths to superintelligence.

Developers test GPT-5: great for reasoning, but still lacking in raw coding skill.
Read More

OpenAI CEO warns of an AI boom bubble, comparing current investor exuberance to the dot-com era and cautioning startup founders that “someone’s gonna get burned.”
Read More

GPT-5 updated after backlash — OpenAI softens tone to sound warmer and more human.
Read More

GPT-5 sparks AGI caution — early reviews say it’s a solid step, not a leap.
Read More

US launches USAi — secure federal hub to test OpenAI, Anthropic, Google, Meta models.
Read More

Google unveils Gemma 3 270M, an ultra-compact open-source AI model that runs directly on devices like smartphones and IoT.
Read More

GPT-5 rollout widens, free ChatGPT users gain GPT-5/mini, while Pro users get GPT-5-pro/Thinking and clarified API pricing tiers.

Read More

US in talks to take stake in Intel to strengthen domestic chipmaking amid delays at Ohio fab.
Read More

AMD CEO Lisa Su targets Nvidia dominance with aggressive growth push despite US-China tensions.
Read More

DeepSeek delays new model launch due to incompatibility with Huawei chips.

Read More