Five minutes. What happened, and what to do about it. All stories from 12 October 2026
OpenAI Dots vs Meta Muse: what each costs, what it can touch, and where you can actually use it
Both are agents that keep working after you log off, but neither is on sale to individuals in Ireland or the UK yet.
AI agent teams cost up to 5 times more than one agent and barely score higher, new tests find
The Decoder· 11 OctThe Decoder reports that the evals company Vals AI tested GPT-6 Sol and Claude Opus 5.5 on its Vibe Code Bench, alone and as teams, at medium and maximum reasoning effort. Teams cost between 1.8 and 5.1 times more than single agents. Only one of four comparisons showed a statistically significant improvement: GPT-6 Sol at medium reasoning, where the team scored 7.3 points higher. At maximum reasoning the team setup gave neither model a real advantage.
- When a vendor sells a multi-agent setup, ask for a side-by-side with one agent on your own task, including what each run costs.
- Set a monthly cap on any usage-billed AI tool, so a team of agents cannot quietly run up the bill.
Source: The Decoder. This is our short summary of their reporting. The paragraphs marked "Our read" and the things to do are SeedFoundry's opinion. The summary is drafted with AI tools. Image: Vals AI, from its own announcement. How we source and credit
These are short summaries of other people's reporting, with our own commentary. Each story names its source and links to the original. How we source and credit