Wait a minute - is it me or is OpenAI rolling out managed Agents feature on the OpenAI Platform?

Can anyone see it as well? 👀

Agent creation, environment creation, templates, and session creation now work. Sessions remain idle and throw an error for now.

OpenAI launches GPT-Live-1 for full-duplex voice agents

OpenAI released GPT-Live-1 in the API, a full-duplex voice model for apps and workflows that listens and speaks at once, supports interruptions, cuts latency, and offers prompt-based control, telephony use, and separate backend pricing.

🗞 #openai @testingcatalog

OpenAI announced ChatGPT for Financial Services.

The feature is available to eligible financial institutions and allows users to connect a wide range of financial data to their ChatGPT conversations.

> ChatGPT for Financial Services can be used with GPT-6 Astra inside ChatGPT Work mode.

> GPT-6 Astra scores 9% higher than GPT-5.6 Sol and 7% higher than Claude Fable 5.1 on the OfficeQA Pro benchmark.

ChatGPT Terminal 👀

Perplexity is working on new Automations feature for Perplexity Computer.

Users will be able to schedule their workflows or invoce them based on a certain trigger.

> "Automate anything with Computer - Run workflows on a custom schedule or when specific conditions are met and get results delivered automatically"

OpenAI launches ChatGPT for Financial Services

OpenAI launched ChatGPT for Financial Services, combining financial data and GPT-6 Astra for research, modeling, and client materials, with enterprise controls, source citations, and integrations shaped by Morgan Stanley and Evercore.

🗞 #openai @testingcatalog

DAILY AI BRIEF 🗞 — Sept 11

OPENAI 🔥:
- GPT-Live-1 is live in the API. Full-duplex voice agents that listen while they speak and can hand work to other models.
- Agents API is in public beta. Managed cloud agents on the Codex harness, plus hosted sandboxes. Platform UI is rolling out too.
- ChatGPT for Financial Services is out for eligible institutions. Work mode plus GPT-6 Astra, with Daloopa, PitchBook, and LSEG data.
- New $200 Pro signups paused. Astra demand is straining capacity. Existing accounts, other plans, and the API stay up.

GOOGLE 🔥:
- Gemini desktop app for Windows is out globally on Windows 10 and 11. Alt + Space overlay for drafts, docs, and media.

META 🔥:
- Shared Muse agents are in development, likely headed for Meta Connect later this month.
- Custom Muse voices are in the works. Prompt a voice in chat, then pick it in voice mode. Working-sounds toggle spotted too.

CURSOR 🔥:
- Projects is in beta. One persistent coordinator thread with subagents, shared memory, scheduled tasks, and PR/Slack watchers.

COGNITION 🔥:
- SWE-2 is out in Devin Desktop and CLI. Post-trained on Kimi-K3. Hits 50.0 on FrontierCode and matches Fable 5.1 at 64% lower cost.
- Free for Pro, Max, and Teams for a month. Devin Voice also shipped, on GPT-Live plus SWE-2.

PERPLEXITY 🔥:
- Automations for Computer are in development. Schedule workflows or fire them on a trigger and get results delivered.

* Used Grok to compose this brief, cherry-picking the news and doing some post-editing.

Muse Agent from Meta already has logic to power "invites"; however, it is not active yet. It seems like those would also arrive with extra free usage!

> "Get free additional usage when you redeem invite code."

A global expansion via invite codes right around Meta Connect would be massive.

Seems like Grok Bot is rolling out on Grok web for Grok Heavy users.

> When available, users can create and interact with their bots directly in the web UI.

> No official announcement has been published yet. This rollout may also be gradual or geo-restricted for now.

This is a "Defender's gap" chart that OpenAI published recently. It shows the gap between defenders' capabilities and attackers' capabilities from a cybersecurity POV. This also translates to a gap between proprietary and open AI models.

What Dario is proposing is closely related:

"Thus, a key part of pacing within democracies is to keep democracies’ AI lead over autocracies as large as possible, to give us the breathing room we need in order to pace effectively."

In other words, Anthropic and OpenAI want to widen the "gap" between what their AI can do and what the rest of the world can do.

This doesn't necessarily mean that they will stop AI development and training.

What this leads to is:
- Anthropic, OpenAI, and other frontier labs will need to work together to make sure that every lab maintains alignment standards.
- These labs will continue using RSI to advance their internal models with "employee-like access".
- These models WON'T be released to the public until the "gap" is sufficient and until alignment standards are met.

What about China?

> "Distillation of frontier models allows lagging companies to narrow the gap using a fraction of the cost it would take to develop their own AI independently."

> "If we execute these measures well, I believe they would slow China’s progress enough to widen America’s lead significantly over the next 3–5 years — the window when AI becomes geopolitically most important."

The assumption behind these measures is simple: without being able to distill frontier models, it will take China significantly longer to close the gap with top-tier models.

All the above may fay fail. China may or may not take the lead in AI progress.

Yet, it's not a surprise that AI can already be used as a cybersecurity weapon. Note that the top point on the "frontier" line describes defenders' capabilities available to companies with Daybreak access and similar. However, the top point on the "open-weight" line is accessible to everyone.

Even with current levels of intelligence, we will start seeing more and more major security incidents around the globe.

> “Everything that makes it successful is exactly what makes it dangerous.”

We should be monitoring this very closely 👀

ICYMI: Cursor announced Projects for agent coordination

Cursor Projects, now in beta, coordinates long-running software work through delegated agents. It runs tasks in cloud or local environments, supports scheduled and PR-driven maintenance, and targets work spanning multiple PRs.

🗞 #cursor @testingcatalog