ChatGPT is the most popular chatbot, by a high margin (44% versus Gemini at 24%). I remain surprised at the high use of Meta AI (14%), but not surprised at the ratio of ChatGPT to Claude (7.33:1). A good reminder at how few people must be using Codex or Claude Code or
AGENTS
-
Time-Lapse Verification of the Strait of Hormuz Status
By
–
Is the Strait of Hormuz open or closed? Came up over lunch so I voice noted my agent to "boot up god's eye view and check."
— Bilawal Sidhu (@bilawalsidhu) 22 juin 2026
It sent me back this timelapse clip — refreshed the AIS data, checked oil futures, mapped 120 days of vessel transits and rendered it out.
TL;DR it's… pic.twitter.com/sTveBbkeBjIs the Strait of Hormuz open or closed? It was brought up at lunch, so I sent a voice message to my agent to 'start the God's eye view mode and check.' He sent me this time-lapse sequence — he updated the AIS data, checked the contracts at
-
GLM-5.2 (max) 3rd on agentic benchmark GDPval-AA
By
–
Absolutely incredible: GLM-5.2 (max) places 3rd overall on GDPval-AA, a real-world agentic work benchmark, even ahead of GPT-5.5 (xhigh). Oh and by the way: it seems open source is no longer 7 months behind. GDPval-AA, a benchmark built around tasks.
-
Live discussion on building, deploying, and managing agentic apps
By
–
Most agentic apps never make it past the prototype stage. Join Databricks co-founder @pwendell
, OpenClaw creator @steipete
, @OpenAI
's Thibault Sottiaux, and @Stellantis
' Head of Agentic AI, Hugo Sechier for a live discussion on what it takes to build, deploy, and manage -
8x code output makes verification the top problem
By
–
My biggest takeaways from Claude Code/Cowork lead @Nerdi_Yogi
: 1. When your engineers ship 8x more code than a year ago (like they do at Anthropic), the biggest problem becomes verification. How do you know that the experience you shipped is what you intended? One tactic Fiona’s -
Bland raises $100M for voice AI handling complex calls
By
–
Most voice AI breaks down the moment a conversation gets complicated.
— AI Highlight (@AIHighlight) 22 juin 2026
Bland raised $100M to build the one that does not. Urgent calls. Long calls. High pressure calls.
That is the entire product. https://t.co/699IKJGE2lMost voice AIs collapse as soon as a conversation becomes complicated. Bland raised $100 million to create one that doesn't. Urgent calls. Long calls. High-pressure calls. That's the whole product.
-
Delusional spirals in chatbots: warning signs and solutions
By
–
When chatbots are programmed to please, some interactions can escalate into what researchers call 'delusional spirals.' Learn more about the warning signs and what can be done: https://
brnw.ch/21x3ysH -
European AI as addressable colleagues with persistent identity
By
–
EUROPEAN AI IS COOKING AGAIN♨️♨️♨️
— Charly Wargnier (@DataChaz) 22 juin 2026
Most AI resets post-task.
I love tools refining long-term memory, and Delos' Workers nail it.
With persistent identity, email, phone & Slack, they sit in the org chart like addressable colleagues.
Curious to see how this autonomy holds up 👀 https://t.co/2JNStMoMQ2EUROPEAN AI IS COOKING AGAIN Most AI resets post-task. I love tools refining long-term memory, and Delos' Workers nail it. With persistent identity, email, phone & Slack, they sit in the org chart like addressable colleagues. Curious to see how this autonomy holds up
-
Interactions API GA: Orchestrate Across Models and Agents
By
–
Today we are shipping the Interactions API into GA, our new API designed to let you orchestrate across models and agents in the same interface, and is our new default API going forward. Interactions sets the stage for the new era of Agents : )
-
OpenAI Daybreak models discover and patch critical vulnerabilities
By
–
We're accelerating patching, in addition to vuln finding, with new tools and models in OpenAI Daybreak. Our models are now discovering and generating patches for critical vulns in major browsers, network infrastructure, and operating systems (such as FreeBSD and the Linux