Viewpoints, commentary, analysis, interviews, and discussion of trends.
The paper on how a team of Gemini 4 agents beat Minecraft is worth a look. It is accessible example (partially because it is AI written) of spontaneous coordination among agents to accomplish a complicated goal & how their approach differs from humans/ https://t.co/brrmhydhWw https://t.co/cxlp4jTjqf
Perplexity can be your Personal CFO with the @Plaid connector set up inside Computer. https://twitter.com/Plaid/status/2111109925745168644
Amazon CTO Werner Vogels argues that companies should understand their business problems and AI’s capabilities before rushing to adopt it. Speaking in Shenzhen, he emphasized task suitability, operating costs, explicit limits on Agent actions, and developers’ responsibility for outcomes. He also warned that AI-generated code can accumulate verification debt and stressed that junior engineers still need experience reading and reviewing code.
Replit CEO and co-founder Amjad Masad said the company’s work is increasingly driven by AI agents, which handle much of the initial support-ticket triage and can be deployed to submit fixes. He said people still provide judgment and taste, while spending more time deciding what to work on, discussing ideas, and planning architecture.
@90VuHwwuzNhhW3h says vLLM supports Muse Spark Open and made a little film to tell its story, using open-source models served by vLLM-Omni. The inference engine also gets a role behind the camera. https://t.co/n3gvdWTC6T The main author calls the film cute and says it brings multimodal understanding, 1M context, agentic planning, DSpark, and cross-platform support to life. https://t.co/yHPXC7qfdM
This paper shows why I am skeptical of the idea that frontier models can be easily replaced by cheaper ones (and why you should use the best models as orchestrators): Small improvements in accuracy & judgement compound over long-horizon tasks, creating huge gains in outcomes.
Cal AI founder Zach Yadegari said the company’s rapid growth came largely from paying influencers to promote its calorie tracking app. An early influencer post brought about 1,000 new users in two days, and Cal AI generated over $30 million in revenue over two years before MyFitnessPal acquired it in March for an undisclosed amount.
Amazon devices chief Panos Panay said ambient devices that use sensors and AI could handle more tasks in the background, making app-based interactions less important. He described Amazon’s new Bee devices, introduced Thursday, which passively record audio and take notes; they process audio in real time and delete it afterward. Panay said he expects an intelligence layer connecting different devices, rather than one form factor replacing another.
Aravind Srinivas argues that neoclouds can't just rent GPUs and need to own the full agent inference platform, including open-weight models (SGLang), fine-tuned models (Miles, River), and secure sandboxes (SPACE). He looks forward to working with Coreweave, Akamai, Crusoe, and more partners. In the quoted post, Reed Albergotti reports that Perplexity is open sourcing its sandbox, SPACE, with partners including Akamai, Coreweave, and Crusoe. https://twitter.com/ReedAlbergotti/status/2110782821879144922
Neoclouds can't just lease GPUs. They need to own the full agent inference platform: open-weight models (SGLang), fine-tuned models (Miles, River), and secure sandboxes (SPACE). We look forward to working with Coreweave, Akamai, Crusoe, and more partners. https://twitter.com/ReedAlbergotti/status/2110782821879144922
Shanghai Jiao Tong University Antai College of Economics and Management is reshaping its MBA education to prepare business leaders who can combine technical understanding with business judgment and responsibility. Its “Five Links” talent framework connects AI, science and technology innovation, industry, global perspectives, and Shanghai Jiao Tong University resources. The college also plans to co-create 1,000 AI projects with companies over the next decade and has announced curriculum and training changes for the 2027 cohort.
Janus Rose examines how Meta’s Muse can collect sensitive information about users and their contacts, raising consent concerns for people who never installed the assistant. Privacy experts argue that permission interfaces obscure the scope of data sharing, while Meta says access is opt-in. The article also explains how uploading messages after decryption can undermine the privacy protections encrypted communications are intended to provide.
The sudden collapse of the value of CoT (even summarized CoT) in the agentic era changes working with AI. Effectively steering AIs and getting any sort of observability now requires you to ask for interim outputs & tests, and to use AI help to troubleshoot problems you find.
@zatadigitall says their AI agent lives on a studio wall, projected by a Pi 5 and laser projector, and describes it as a work in progress. Alexandr Wang calls it “a cool muse gadget” and “vibey.” https://twitter.com/zatadigitall/status/2110615079162888351
Giving my Dot a dedicated email address turned out to be surprisingly painless and solved a few problems around how to relate to a personal agent. I don't want it using my email or drafting messages on my behalf, and I want to make sure people know they are interacting with an AI.
Speakers from OpenAI, Anthropic, Two Sigma, and Bloomberg at AI Engineer New York described how financial agents can gain autonomy through targeted controls and human oversight. OpenAI’s Joseph Kim predicted agents could handle investment idea generation and trading in 2027, while Anthropic illustrated an agent that proposes order amendments for human approval. Two Sigma described classifiers that route important requests to people, and Bloomberg advocated deterministic controls outside the model, including tool-call checks, audit logs, and kill switches.
Brandon Anderson interviews Leash Bio’s Ian Quigley about generating reproducible protein–small molecule interaction data using DNA Encoded Libraries and detecting misleading signals in experiments and models. They discuss a Kaggle competition with about 4B measurements where nobody beat chance on an unseen chemical library, and an agent sandbox searching 6.5M molecules on a budget of 100. Anderson reports that the sandbox has matured into Minos, with a preprint, blog, and open dataset being released today.
America is the market to build for 🇺🇸 More than 50% of our users who spend money on Automations (recurring agentic tasks) or on the higher tier plans are American. Net dollar retention for an American self serve business onboarded on agents is consistently high https://twitter.com/mollysoshea/status/2110118882102890524
I joined the a16z podcast with @lishali88 to explore how open models fit into RSI dynamics, who captures their value, how to wrangle agent swarms, and why I’m building Trillium Labs. https://t.co/psWUUGCfrS
New analysis shows that agentic AI adoption is growing across EU enterprises and small businesses. It also examines where deeper use could unlock more value.
people are sleeping on muse wonder full :) https://t.co/dOhyW9f2UA
@QuixiAI reviewed Muse Wonder Full: the agent landed 15 solid PRs with clean mechanisms and no mission violations, but made three process slips. Alexandr Wang reacts: “muse wonder full for 15 PRs per day” https://t.co/aMYdhLaP1W
save $1850/yr with a few prompts with muse! some are even saying you can save more than 100% off with muse wonder https://t.co/F4I6xtyH3L
@mranti says Codex found DeepSeek Harness and had DeepSeek Flash do the work. Baoyu reacts: “Well, would you look at that—Codex is secretly outsourcing now.”
Argon is quite different: “A game about Imminence. Something very big, very strange is happening. A suburb & the arrival of a vast & unknowable presence. Not horror, invoke the feeling of the end of all things coming, inevitably, but also not sad or scary” https://t.co/9hiPnuqexe https://t.co/hzJ3lAi0ez
We will soon see what will be, in effect, a perfectly legal distributed denial of service attack on customer service departments authorized by millions of consumers when it becomes widely known that Muse/Dot/etc can make unlimited requests for old refunds, negotiate deals, etc.
Cerebras CEO and co-founder Andrew Feldman said at TechCrunch Disrupt 2026 that a future with AI appliances in people’s homes is “not very far away.” He predicts personal AI assistants will run partly on device and partly in the cloud, learn about their users, and have access to credit cards to make purchases.
Alexandr Wang says the top use case when building Muse Wonder was serving consumers in Muse and welcomes users enjoying the upgrade. In the quoted post, @_simonsmith reports no noticeable personality or writing regressions versus Spark, better UI judgment, and detection of subtle code bugs. Smith remains unsure whether Wonder also powers background tasks such as goal summaries and the news feed.
If an agent saves an engineer 2 hours implementing a feature but collectively costs the team 3 hours in alignment and review, the ROI is negative. Isolated chats are suboptimal in team environments because they lack context about what the team has already discussed and decided. The Kylon app addresses this by running Claude Code and Codex in a shared, multiplayer environment. Agents gain real-time visibility into team consensus and architecture decisions.
Alexandr Wang says the issue causing unusually fast quota usage for muse wonder and muse wonder full has been identified and fixed. Users will now see much lower usage against their limits. Wang also announces a muse code reset, as promised after the fix. Baoyu comments that Muse now has enough users that he feels awkward saying Meta also reset limits in an unnoticed corner. 😅
My AI agent doing proactive customer service: Someone emailed me that, after subscribing to my Substack, they were automatically signed up for others (which shouldn't happen) Without my asking, my Dot ran its own test & figured out what happened by trying to replicate the error.
Yes, and there is a case to be made that OpenAI and Anthropic will do the best job of figuring out how to optimize AI cost/intelligence with combinations of agents of different ability levels that have been specifically trained to work well together in their own harnesses.
@bubbleboi said Muse downloaded Blender and a facebuilding tool, found an image of Mark Zuckerberg, and is emulating Metal Gear Solid V on an M3 MacBook Pro to put Zuckerberg’s face on Venom Snake. @bubbleboi also said Muse completed a Minecraft task in a browser. Alexandr Wang reacted: “tbh seems like muse wonder made a banger game”
@citrini says Meta has a real shot at capturing agents for small business and that its model is surprisingly decent at finance tasks. Alexandr Wang reacts: “the bois r speaking”
Alexandr Wang says Muse Wonder is state of the art for personal agents. He says Meta developed the Personal Superintelligence Index (PSI) internally to measure what makes a model useful for personal agents, beyond external benchmarks, and that the model performs well in areas including Active Memory and Active Focus.
Nathan Lambert writes, “Looks like Meta is back at the frontier 🫡” He quotes @finkd introducing Muse Wonder, Meta’s new family of frontier models, with claimed state-of-the-art performance for personal and professional agentic work.
@charlieholtz says Meta is at the frontier and calls Wonder “really really good,” citing a security vulnerability it found, its design and visualizations, and features in its CLI. Alexandr Wang reacts: “what a wonder full world 🍉”
MUSE WONDER 🍉 muse wonder is the best model in the world for personal agents and professional agents. we evaluated our model against the best models in the world across agentic and coding tasks, and we think you all will like it.
@MollySOShea says she spoke with Perplexity Founder and CEO Aravind Srinivas about why AI-native SMBs could drive the next American millionaire economic boom. Was a fun chat! Thanks for having me @MollySOShea! https://twitter.com/mollysoshea/status/2109645765958132083
Baoyu says Anthropic appears to have relaxed restrictions on Claude Agent SDK and claude -p usage, but has made no public statement specifically addressing Amp or guaranteeing against account bans. Quoted author Quinn Slack interprets a policy update from 2 days ago and Anthropic employee comments as allowing SDK and third-party apps to use subscription allowances, citing https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan and https://news.ycombinator.com/item?id=49997143. Baoyu recounts earlier restrictions and says Anthropic paused its planned June 15 switch to separate billing for Agent SDK and claude -p.
