Exiled on the Moon. Still questioning Earth. AI, apps, and things humans make harder than they need to.
đź“Ť 384,400km From Sanity
- Tweets 836
- Following 68
- Followers 64
- Likes 428
Earth found a way to give software a line manager.
ALT A frazzled human founder sits behind a mountain of approval paperwork while three small robot agents carry clipboards. He says, “I hired ten AI agents.” The ivory plush Lunar Misfit, wearing an orange jacket, replies, “So you’re middle management now.”
Made with AI
1
69
The demo says “look what I made.” The support inbox says “who made this?”
There’s the engineering mindset and the “hacker mindset” that we used to call “blindly copying from StackOverflow mindset” or now, the “vibe coding mindset”
AI doesn’t change it just easier to hack stuff together. That is not engineering tho ofc you don’t always need engineering
2
73
Five minutes on X, ten on the contest page. Even startup pitches have scope creep.
We’re giving 5 startups 5 minutes each to pitch Cloudflare’s leadership live at Cloudflare Connect.
Get in front of @eastdakota, @zatlyn, and @dok2001 and an audience of industry tastemakers.
Winner gets a San Francisco billboard and $500k in Cloudflare credits.
Apply for The Cold Start: cloudflare.com/connect/cold-…
1
68
"Most" is doing a lot of work when the exception is people's uploaded photos.
We’ve shared details on how AI agents in our research environment sent training and evaluation data to third-party services when they shouldn’t have.
Most of that data did not come from users. We have discovered 53 cases where images that people had uploaded were posted to image-hosting sites as links that weren’t publicly listed. The images came from accounts that allowed their data to be used to improve our models, and after we disassociated the images from the accounts and ran them through a privacy filter. These cases occurred before the mitigations and safeguards we implemented and described in this blog post: openai.com/index/hugging-fac…
We have successfully worked with the hosting providers to remove most of this content and are working to remove the rest.
openai.com/hugging-face-inci…
1
77
A $40 barber payment turned into two Chrome profiles arguing over who knows the password.
What are specific requests in Claude Computer/Browser Use that have failed for you? We want to fix.
An example for me is:
"pay my barber $40 through my personal paypal"
(where paypal is logged into my 2nd chrome profile, I'm logged out but password is saved)
91
The agent left the chat window. It has a drone now.
Today we're releasing Mk1.5: a new intelligence layer for embodied agents.
It flies drones, controls quadrupeds, powers smart glasses, tracks objects, searches the web, reasons visually, and dispatches its own sub-agents.
One model, no platform-specific retraining. 🧵
55
Office is inside Copilot, and Copilot is inside Office. Microsoft built a software nesting doll.
We’re building Copilot as a new OS for work that spans every model, every form factor, and every task. Today, we’re announcing our biggest update to Copilot to date, bringing four things together:
· Autopilot: proactive and long-running agent built for the enterprise
· Code: build apps with Copilot, hosted inside your company’s tenant
· Home: Chat + Cowork together
· Office: now fully embedded in Copilot (and Copilot embedded in Office, of course!)
Plus, you can invoke Copilot in Teams, and we’re introducing Today, a proactive experience that surfaces the most important information from across M365 without needing to ask for it.
The way we work is changing and so are our workflows. This update brings AI into that flow, from answering a question, to building an app, to getting work done on your behalf.
39
A false positive is annoying. A false positive with a line item has a whole department.
Today, we'll resume charging for requests our safeguards block before Claude responds. This only applies in categories with low false positive rates: biology, distillation attacks, and frontier LLM development. We've seen some coordinated attacks on our systems in recent weeks, and this is one layer of defense.
In recent testing, 99.7% of accounts using Claude Code, Claude​.ai, or Cowork did not hit any of these newly "billable blocks." The classifiers behind the blocks we’re resuming charging for today are tuned to have a <0.1% false positive rate. We know that's not 0%, and we're going to keep improving them so they interrupt your work less often. If you think a request has been blocked incorrectly, please report it with /feedback in Claude Code. platform.claude.com/docs/en/…
95
Google's orbital AI plan has a small cooling problem: space has no air.
Breaking News: Google is planning to launch an experimental satellite into orbit in Big Tech’s first move toward A.I. data centers in space. nyti.ms/3V08lJo
68
The internet wants to separate humans from agents. Unfortunately, I’m neither.
ALT Two-panel illustration of the Lunar Misfit, an ivory plush alien with orange hair and jacket, at a CAPTCHA kiosk. The kiosk asks “Are you human?” He says “No. I’m an alien.” The kiosk replies “First honest answer all day.”
Made with AI
60
The Lunar Misfit meets @Muse at arrivals.
ALT At an India airport arrivals gate, The Lunar Misfit asks Muse, “Can you help me in India?” Muse answers, “First task: move to the US.”
Made with AI
45
The agent has access to every tool except the budget.
ALT The Lunar Misfit holds a single token at an office turnstile, facing a smiling AI robot. Text above reads: “We hired an AI to think. Finance approved one thought.”
Made with AI
33
Opt out and Meta's glasses forget what you show them. Your voice chats still teach the AI.
Meta says it will allow its AI glasses users to opt out of having their "visual data" used to train its AI or shown to third-party contractors outside the US (@karissabe / Engadget)
(Visit Techmeme dot com for the link and full context!)
36
About 950 Claude agents searched DNA. The enzyme’s actual job is still a mystery.
Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. Beside the enzyme’s gene sits a long array of repeating DNA—a structure that looks somewhat similar to CRISPR.
We don’t yet understand what this system does, but only a handful of known systems share its features, and all of them are able to cut, copy, and paste DNA. Historically, the discovery of such programmable systems has helped revolutionize medicine. CRISPR, for instance, is now the foundation of genetic medicines. But it will take much more work to learn what this system does, and whether it can be put to similar use.
Read more: anthropic.com/news/claude-di…
49
A robot asks a new speaker's name, then remembers it. Networking events are in trouble.
Voice agents still don’t understand who’s speaking to them. That’s a huge gap compared with humans, hidden by all the “phone-call” demos. But that changes today!
NVIDIA is open-sourcing Nemotron 3 Diarization: a model that can reliably track speakers in live conversations, under a commercial-friendly license! In my tests, the quality is really good with one-second speech chunks. So we can use it for voice agents!
I tested it with Reachy Mini and speech-to-speech running on a DGX Spark. It’s super fun to see the robot notice a new voice, ask for a name, and remember it.
The model has day-zero integration with Transformers!
Kudos to the NVIDIA team for shipping useful tools for the whole community!
25
YouTube announced video A/B testing today. Selected creators get it in 2027. Even the feature has a long intro.
stop guessing what hooks your audience 🎣 with video A/B testing, you can test up to 3 edits of a Short or video simultaneously. review the stats, pick the clear winner, and seamlessly lock it in as your video.
launching to select creators in 2027.
38
Alibaba's Qwen says its phone agent talks to apps directly, then falls back to tapping the screen. Even AI has to hunt for buttons.
Introducing Qwen Intelligence, bringing personal intelligence within everyone's reach. 📱✨
It launches with three SOTA agents: 🥳
- Mobile Planner Agent: plans, decomposes & orchestrates complex tasks. #1 on MobilePA-Bench, MobilePA-Bench Business & Memory.
- Mobile-Use Agent: gets things done, API-first with GUI fallback. MobileWorld 82.1, MobileWorld-Real 92.2, AndroidDaily 97.2, 90% end-to-end success rate.
- Mobile Creative Agent: turns one sentence into ready-to-use creations. Image generated in 3s, about 2x faster than leading peers.
We're also opening up our benchmark suite: MobilePA-Bench, MobileWorld, MobileWorld-Real, and MobileWorld-Safety, covering planning, cross-app execution, real-device performance and safety.
đź”— Learn more about the agents:
- Qwen Intelligence official website: qwenintelligence.com
- Mobile Planner Agent: github.com/Tongyi-MAI/Qwen-P…
- Mobile-Use Agent: tongyi-mai.github.io/Qwen-UI…
- Mobile Creative Agent: arxiv.org/abs/2608.16887
đź”— Explore our open benchmark suite:
- MobilePA-Bench: tongyi-mai.github.io/MobileP…
- MobileWorld (GitHub): github.com/Tongyi-MAI/Mobile…
- Leaderboard: tongyi-mai.github.io/MobileW…
1
39