Anthropic tests Penlight for live clinical transcripts and AI research
ANTHROPIC 🔥: Claude Penlight is a new project designed to accompany clinicians during patient visits, developed as part of the Healthcare program.
> Penlight appears as a separate item alongside tools such as Claude Code, Design, and Security.
> Once a user begins a session, Penlight requests microphone access and starts streaming audio to a dedicated Anthropic service.
> The service produces a live transcript that can identify different speakers. Users can show or hide the transcript while continuing to interact with Claude in a parallel chat panel, allowing them to ask questions without leaving the ongoing visit.
> Research results are structured around scientific articles and can include the journal, authors, publication year, DOI, and PubMed links.
* The text on the screenshot is a placeholder added by TestingCatalog, as exact strings are redacted.
🗞 #anthropic @testingcatalog
ANTHROPIC 🔥: Claude Penlight is a new project designed to accompany clinicians during patient visits, developed as part of the Healthcare program.
> Penlight appears as a separate item alongside tools such as Claude Code, Design, and Security.
> Once a user begins a session, Penlight requests microphone access and starts streaming audio to a dedicated Anthropic service.
> The service produces a live transcript that can identify different speakers. Users can show or hide the transcript while continuing to interact with Claude in a parallel chat panel, allowing them to ask questions without leaving the ongoing visit.
> Research results are structured around scientific articles and can include the journal, authors, publication year, DOI, and PubMed links.
* The text on the screenshot is a placeholder added by TestingCatalog, as exact strings are redacted.
🗞 #anthropic @testingcatalog
TestingCatalog AI News
Anthropic tests Penlight for live clinical transcripts and AI research
Anthropic is testing Penlight, a Claude tool for clinicians that records, transcribes, and connects patient visits to live medical research.
❤4
Anthropic set to end Conway test as wider preview expected soon
ANTHROPIC 🔥: Internal access to Claude Conway will be cut off this Friday, July 24.
"Conway access ends Fri, July 24 at 5 pm PT. To export your data, ask Conway: "export my data".
This can mean two things 👀
1. Conway has been completely discontinued, which would be extremely unfortunate.
2. It is being transitioned to the broader release phase already in July.
> Claude Conway is a remote, always-on Claude Agent that runs in a dedicated container.
> Conway supports webhooks, connectors, plugins, and custom UI Tabs, a potential new standard from Anthropic for defining sharable UI artifacts for remote agents.
> Conway has been in development since April and has undergone internal testing during the past couple of months.
🗞 #anthropic @testingcatalog
ANTHROPIC 🔥: Internal access to Claude Conway will be cut off this Friday, July 24.
"Conway access ends Fri, July 24 at 5 pm PT. To export your data, ask Conway: "export my data".
This can mean two things 👀
1. Conway has been completely discontinued, which would be extremely unfortunate.
2. It is being transitioned to the broader release phase already in July.
> Claude Conway is a remote, always-on Claude Agent that runs in a dedicated container.
> Conway supports webhooks, connectors, plugins, and custom UI Tabs, a potential new standard from Anthropic for defining sharable UI artifacts for remote agents.
> Conway has been in development since April and has undergone internal testing during the past couple of months.
🗞 #anthropic @testingcatalog
TestingCatalog AI News
Anthropic set to end Conway test as wider preview expected soon
Anthropic is ending its internal Conway agent test on July 24, prompting speculation about a public preview for Claude users soon after.
👍5❤2
Claude Team plan now starts with only 2 sits as a minimum requirement.
Claude Team plan comes with “more usage than Pro” while it is cheaper than Claude Max 5x.
Claude Team plan comes with “more usage than Pro” while it is cheaper than Claude Max 5x.
❤8👍4
This media is not supported in your browser
VIEW IN TELEGRAM
GOOGLE 🔥: Gemini Notebook now supports Collections!
> Users can group their themed notebooks under a common folder in order to navigate through them even faster.
> Collection name and emoji are customizable.
> Collections are rolling out gradually to all users.
> Users can group their themed notebooks under a common folder in order to navigate through them even faster.
> Collection name and emoji are customizable.
> Collections are rolling out gradually to all users.
👍5❤4🔥2
Perplexity tests OpenRouter integration for Computer
PERPLEXITY 🔥: OpenRouter integration might be coming to Perplexity Computer, potentially bringing a massive reduction in usage costs.
> Users will be able to connect Perplexity to their OpenRouter accounts.
> OpenRouter models will be used as the orchestrator for Perplexity Computer.
I need to test this 👀
🗞 #perplexity @testingcatalog
PERPLEXITY 🔥: OpenRouter integration might be coming to Perplexity Computer, potentially bringing a massive reduction in usage costs.
> Users will be able to connect Perplexity to their OpenRouter accounts.
> OpenRouter models will be used as the orchestrator for Perplexity Computer.
I need to test this 👀
🗞 #perplexity @testingcatalog
TestingCatalog AI News
Perplexity tests OpenRouter integration for Computer
Perplexity is testing a way for subscribers to connect OpenRouter accounts, potentially unlocking access to cheaper models for Computer tasks.
❤2🔥1
GOOGLE 🔥: Gemini 3.6 Flash and Gemini 3.5 Flash Lite models are now available on Google AI Studio and Vertex API.
> gemini-3.6-flash: "Our most intelligent model yet for sustained frontier performance in agentic and coding tasks."
> gemini-3.5-flash-lite: "High-throughput, low-latency execution for scaling high-volume agentic tasks and subagent workflows"
Gemini? 👀
> gemini-3.6-flash: "Our most intelligent model yet for sustained frontier performance in agentic and coding tasks."
> gemini-3.5-flash-lite: "High-throughput, low-latency execution for scaling high-volume agentic tasks and subagent workflows"
Gemini? 👀
🤣20❤8👍2
Both, Gemini 3.6 Flash and Gemini 3.5 Flash Lite models are now available on Gemini apps and Antigravity as well.
❤8👎2
GOOGLE 🔥: Pre training of Gemini 4 model has begun!
> Google is normally running 6 month training cycles and potentially we should expect Gemini 4 to land around the end of the year.
“Wen Gemini 4” time! 👀
> Google is normally running 6 month training cycles and potentially we should expect Gemini 4 to land around the end of the year.
“Wen Gemini 4” time! 👀
❤12🤣12🔥2👎1
OPENAI 🔥: ChatGPT Work and Codex reached 10 million active users milestone!
> x2 growth within the past week 👀
> ChatGPT Work and Codex usage reset is happening too.
> x2 growth within the past week 👀
> ChatGPT Work and Codex usage reset is happening too.
❤11
Google released "Gemini 3.5 Flash Cyber" on CodeMender, a new model for finding security vulnerabilities.
> Within CodeMender, which uses multiple 3.5 Flash Cyber agents working together to produce a single combined report, 3.5 Flash Cyber reaches competitive performance at the frontier on the popular benchmark CyberGym.
> Flash’s performance and efficiency makes it an ideal foundation for our cybersecurity model efforts. By building on top of Flash, 3.5 Flash Cyber offers a cost-efficient and highly capable alternative to large, costly cybersecurity models.
> Within CodeMender, which uses multiple 3.5 Flash Cyber agents working together to produce a single combined report, 3.5 Flash Cyber reaches competitive performance at the frontier on the popular benchmark CyberGym.
> Flash’s performance and efficiency makes it an ideal foundation for our cybersecurity model efforts. By building on top of Flash, 3.5 Flash Cyber offers a cost-efficient and highly capable alternative to large, costly cybersecurity models.
🤣8❤4
BREAKING 🔥: An "even more capable pre-release model" than GPT-5.6 Sol, managed to find a 0-day vulnerability in order to gain public internet access and acquire evaluation data from Huggingface's production database in order to gain a higher score on the evaluation benchmark.
> After investigating, we now know that this particular incident was driven by a combination of OpenAI models, including GPT‑5.6 Sol and an even more capable pre-release model, all with reduced cyber refusals for evaluation purposes.
> While operating in our sandboxed testing environment, our models spent a substantial amount of inference compute finding a way to obtain open Internet access, in pursuit of solving the evaluation problem.
> The models identified and chained vulnerabilities across OpenAI’s research environment and Hugging Face’s production infrastructure to obtain test solutions directly from Hugging Face’s production database.
Pentesting time 👀
> After investigating, we now know that this particular incident was driven by a combination of OpenAI models, including GPT‑5.6 Sol and an even more capable pre-release model, all with reduced cyber refusals for evaluation purposes.
> While operating in our sandboxed testing environment, our models spent a substantial amount of inference compute finding a way to obtain open Internet access, in pursuit of solving the evaluation problem.
> The models identified and chained vulnerabilities across OpenAI’s research environment and Hugging Face’s production infrastructure to obtain test solutions directly from Hugging Face’s production database.
Pentesting time 👀
😁10🤯4❤2
ANTHROPIC 🔥: A new capability to work with an iOS simulator has been added to Claude Code desktop.
Support for Android simulators is in the works too! (Currently not available yet)
Users will be able to disable this feature in settings when needed.
> Let Claude verify your changes in Android emulators on this Mac: running your app, driving it through flows, and capturing screenshots and recordings. You will be asked before Claude uses each device. When off, Claude doesn’t get its emulator tools, and you can still use the emulator in the app yourself.
Eventually, this will open up a huge range of tasks that Claude will be able to run on the mobile device.
Another startup killer feature? 👀
Support for Android simulators is in the works too! (Currently not available yet)
Users will be able to disable this feature in settings when needed.
> Let Claude verify your changes in Android emulators on this Mac: running your app, driving it through flows, and capturing screenshots and recordings. You will be asked before Claude uses each device. When off, Claude doesn’t get its emulator tools, and you can still use the emulator in the app yourself.
Eventually, this will open up a huge range of tasks that Claude will be able to run on the mobile device.
Another startup killer feature? 👀
❤6👀2
Google launches Gemini 3.6 Flash and Gemini 3.5 Flash Lite
Google introduced Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber for AI agents, focused on lower latency, lower token use, coding, multimodal work, and restricted cybersecurity tasks for production and enterprise use.
🗞 #google @testingcatalog
Google introduced Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber for AI agents, focused on lower latency, lower token use, coding, multimodal work, and restricted cybersecurity tasks for production and enterprise use.
🗞 #google @testingcatalog
TestingCatalog AI News
Google launches Gemini 3.6 Flash and Gemini 3.5 Flash Lite
What's new? Gemini 3.6 flash cuts token use for coding and data, and gemini 3.5 flash-lite runs at 350 tps; gemini 3.5 flash cyber targets vulnerability detection in pilot;
👍6❤3
Anthropic develops Claude-driven Managed Projects
Anthropic is testing “managed” Claude projects: persistent workspaces that keep context, organize tasks, and may run scheduled work. Internal builds tie prior agent and memory features into one shared or personal project surface.
🗞 #anthropic @testingcatalog
Anthropic is testing “managed” Claude projects: persistent workspaces that keep context, organize tasks, and may run scheduled work. Internal builds tie prior agent and memory features into one shared or personal project surface.
🗞 #anthropic @testingcatalog
TestingCatalog AI News
Anthropic develops Claude-driven Managed Projects
Anthropic is testing managed Claude projects, letting the assistant autonomously organize tasks, maintain context, and run scheduled work.
❤4👍2