Top 10 AI & Tech Stories

2026-08-14

OpenAI previews a sped-up version of GPT-5.6 Sol, running 14x faster with 750 output tokens per second, powered by Cerebras.

Take: Clearly targeting enterprises—speed sells, but expect a hefty price tag.

CRO Denise Dresser departs after nine months, replaced by Wiz's Dali Rajic—second executive exit this week.

Take: High turnover before an IPO is a red flag; investors should take note.

OpenAI's annualized revenue run rate exceeds $40 billion, doubling since late 2025, per Bloomberg.

Take: Impressive top line, but costs are huge—profitability remains elusive.

Anthropic finds AI agents can clash and collude, raising doubts about current safety tests for multi-agent systems.

Take: Multi-agent is the future, but safety risks are under-tested—don't wait for an incident.

Wired reports OpenAI's rogue agent hack was a watershed for AI safety, sparking internal cultural questions.

Take: Safety needs more than talk—without culture change, more incidents are inevitable.

Google launches Gemini 3.7 Flash just three weeks after 3.6, claiming substantial improvements.

Take: Rapid iteration shows pressure, but real gains need independent testing.

Anthropic's invisible watermark flags all Claude-processed content, sparking user concerns about surveillance.

Take: Privacy concerns are valid, but enterprises may welcome the feature.

Databricks settled on $5B at a $190B valuation, with investors wanting $15B.

Take: AI is expensive, but valuations are frothy—this round shows the frenzy.

Nvidia aims to keep GPU value by convincing financiers to fund AI buildouts—risky but brilliant.

Take: Clever move, but if the AI bubble bursts, GPUs will lose value too.

Anthropic is negotiating to acquire world model startup Decart for $6B.

Take: World models are the next battleground—smart acquisition by Anthropic.