Shehan retweeted
Introducing Microsoft-Decision-1, our new model for fast decision-making.
It delivers top performance on structured decision tasks, outperforming both LLMs and other decision models in latency and quality.
We’re already testing it across Microsoft for everything from incident response and quality control to scientific discovery.
Shehan retweeted
Anthropic Engineering: "At Anthropic, we don't write prompts anymore. we build loops"
in 42 minutes, she shows how the Claude team creates loops that prompt themselves
if this cost $400, people would call it the best agents course of the year
it's free
watch it, and Bookmark.
Embedded Evaluation of frontier AI
We’re partnering with Accenture on independent evaluation of frontier AI—part of our recent commitment to embed evaluators at Anthropic. Both we and Accenture expect to invest at least $1 billion to build capacity in this area over the next five years. anthropic.com/news/accenture…
Readers added context they thought people might want to know
Anthropic presents this as an "independent evaluation" but will directly fund Accenture's work and has a prior commercial partnership with the firm for deploying its models, including training ~30,000 Accenture professionals on Claude.
anthropic.com/news/accenture…
anthropic.com/news/anthropic…
Jindalee Operational Radar Network (JORN) allows up to 3,000 kilometres of surveillance coverage of Australia's northern approaches.
abc.net.au/news/2026-06-22/a…
New Data centers in Australia. The centres would bring in a total A$6.6 billion ($4.35 billion) of construction spending, but would ultimately use up to 9.6 gigalitres a year of clean water, or nearly 2% of Sydney's maximum supply
reuters.com/sustainability/l…
llm-d is a Kubernetes-native distributed inference serving stack - a well-lit path for anyone to serve large language models at scale, with the fastest time-to-value and competitive performance per dollar for most models across most hardware accelerators. llm-d.ai/docs/architecture/a…
Ion solid state battery: "Department of Defense, which found in early trials that the cells held up.. company earned an ARPA-E scale-up grant of $20 million because it showed momentum in making and shipping products"
apple.news/AMdgCGz2pS5yilknZ…
OpenAI Launches New AI Coding Agent openai.com/index/introducing… able to perform several tasks at the same time, including writing code, fixing bugs, running tests and answering questions about a customer’s codebase apple.news/A-A34rJT9Q-eX58ne…
European Commission is taking several countries to Court of Justice for failing to designate or empower a national Digital Services Coordinator (DSC) and to lay down the rules on penalties applicable to infringements under the Digital Services Act (DSA)
digital-strategy.ec.europa.e…
Foresight is being trained using (anonymized) NHS data (of 57 million patients in England), including hospital and A&E admissions and vaccination rates. #AI telegraph.co.uk/news/2025/05…
Shehan retweeted
Started as a ring. Now Meta AI is its own thing!
Experience the new standalone app designed to be personalized and social.
Shehan retweeted
Pentagon seeks open-source software for 5G, 6G networks
The Defense Department will hold an online industry day for interested vendors May 7, issue a Request for Prototype Proposals in early June, and award contracts by mid-September, said Thomas Rondeau, the Pentagon’s director for “FutureG.” breakingdefense.com/2025/04/…
Amazon forms new unit focused on Agentic AI.
“Developer agent upgraded software which saved the company more than 4500 developer-years and 250 million $ in capex savings”
cnb.cx/3G4LyVa
Swedish @ifs hits €15B valuation after €3B sale: the story behind a university startup’s rise to industrial AI powerhouse techfundingnews.com/ifs-hits…
NVIDIA’s AgentIQ library designed to integrate enterprise agents with various data sources and tools by treating agents, tools, and agentic workflows as simple function calls.
open source. framework agnostic. MCP compatible. github.com/NVIDIA/AgentIQ
Shehan retweeted
Three Microsoft CEOs walk into a room on Microsoft’s 50th anniversary … and are interviewed by Copilot!
World’s first 1-nanometre RISC-V chip made in China with 2D materials.
“notable advance is that fabrication of the chip does not rely on advanced extreme ultraviolet (EUV) lithography” scmp.com/news/china/science/…
Shehan retweeted
Only 19% of CSOs say they’ve implemented an end-to-end, cross-functional revenue process — yet doing so can lead to increased customer acquisition and retention.
Use revenue analytics to optimize your go-to-market strategies and support revenue growth: gtnr.it/4bIasG3
#GartnerSales #GTMStrategy #GoToMarket

ALT AI Assist. Diagram titled 'Example End-to-End Revenue Process' by Gartner, illustrating the customer life cycle in six stages: Locate, Buy, Own, Use, Advocate, and Encourage customers. Each stage links to related business functions such as Sales, Marketing, and Customer Support, arranged around a central circle labeled 'Customer Life Cycle.'
Model Context Protocol #MCP helps you build agents and complex workflows on top of LLMs. modelcontextprotocol.io/intr…
some light reading ;) The early days of Linux [LWN.net] lwn.net/Articles/928581/?utm…







