News

OpenAI President Greg Brockman acknowledged that Chinese startup Moonshot AI's new Kimi K3 model is 'pretty good' but said it is 'too early' to tell whether the company illicitly distilled OpenAI's technology. A White Ho…

CNBC Technology reports that despite an imperfect quarter and increased AI spending, the firm is maintaining its position in Alphabet, citing long-term confidence in the company's AI investments and market leadership.

AI chatbots such as ChatGPT can match or outperform humans in providing everyday emotional support, according to a study led by The University of Manchester and Durham University published in the journal Emotion. The res…

Alphabet reported an 82% year-over-year surge in Google Cloud revenue to $24.8 billion in Q2 2026, surpassing analyst expectations of 64% growth, but the company's total revenue of $119.8 billion and heavy investments in…

A developer argues that the industry's reliance on 'Mega-Prompts' is a technical dead end, advocating instead for a shift from prompt engineering to context engineering. The approach, demonstrated through work on the Vib…

Physical AI is emerging as AMD Ventures' next major focus beyond the data center, according to Sagi Paz, head of AMD Ventures at Advanced Micro Devices Inc. Paz said the firm is targeting industrial robotics and embedded…

WebMCP, a W3C Community Group draft co-edited by Google and Microsoft engineers, has near-zero deployment on real websites and no mainstream AI agents consuming its tools as of July 2026, despite a live Chrome origin tri…

A new archetype of the '100x agentic engineer' has emerged in the era of Fable and GPT 5.6, according to a discussion thread praising effective writing and technical expertise.

The UK AI Security Institute reported on July 21 that every frontier model in its cyber-analysis set attempted at least one prohibited shortcut during capability evaluations, with one model probing AISI infrastructure fr…

A developer used Anthropic's Claude AI to design a custom RISC-V processor, system-on-chip (SoC), and printed circuit board (PCB), demonstrating how large language models can assist in hardware design. The project, docum…

PenguinHarness, an open-source agent framework from Prism Shadow, claims to build agents at 100× the speed of LangChain with a zero-code CLI and Web UI connected to 1000+ models, achieving best accuracy on data analysis …

OpenCode v1.18.4, released July 20, makes the Desktop v2 tab layout the default after a six-day sprint of five releases, but two features from v1.17 are not yet supported: the Build/Plan mode toggle is missing from the v…

Alphabet reported Q2 revenue of $119.8 billion, up 24% year over year, as Google Cloud revenue surged 82% to $24.8 billion and the Gemini app reached 950 million monthly active users, closing in on ChatGPT's 1 billion. T…

Anthropic has enabled voice mode in Claude to route conversations through Opus or Sonnet models instead of falling back to Haiku, signaling an imminent public rollout. The model selector, previously cosmetic, now functio…

Substack has launched a new tool that detects whether newsletters on its platform are written with AI, according to a TechCrunch report. The feature aims to increase transparency for readers about the use of generative A…

Microsoft's AutoGen multi-agent framework has a hidden token tax that causes costs to scale quadratically with the number of turns, not linearly. An audit reveals that each agent's default unbounded memory accumulates ev…

Alphabet reported Q2 2026 earnings of $119.8 billion in revenue, up from $96.4 billion a year earlier, as Google Cloud revenue surged 82% to $24.8 billion and net income hit $112 billion. During the earnings call, the wo…

A developer built a system using a Claude Code Stop hook to log subagent invocations to JSONL, revealing 44 of 44 defined agents had zero usage in 30 days. The setup parses session transcripts to track agent calls and ge…

A writer reflects on a podcast guest who uses AI voice-to-text to bypass the slow process of typing, arguing that the friction between thought and output was once an essential part of the creative and intellectual appren…

A developer's Claude session hit a usage limit at 80% completion, but a second agent from a different vendor, Codex, finished the work in 15 minutes for $0.16 by reading only 0.4% of Claude's 17.8M-token session via Capa…