OpenAI’S GPT-Realtime Released!

Dear Readers,

The speed at which AI is changing the way we communicate is breathtaking. Yesterday we were still writing messages to bots, today we are talking to them in real time - with pauses, intonation, and spontaneous changes in language. This development raises not only the question of how we talk to machines, but also how much language itself will become the interface of the future.

This issue is all about that: the launch of GPT-realtime and its implications for everyday life and the world of work, political power games surrounding computing resources in the UK and the US, and the growing role of security structures at companies such as Anthropic. It also includes figures, voices, and outlooks that show where the industry is heading. Curious? Then it's worth reading on.

In Today’s Issue:

All the best,

OpenAI has introduced a milestone with GPT-realtime: a language model that processes speech directly into speech - without the detour via text. This results in dialogues that are so fast and natural that they are almost indistinguishable from human conversation. Responses come in milliseconds, with intonation, pauses, and even spontaneous language changes.

For the AI community, this represents a new level of real-time interaction. Developers can integrate voice agents directly into apps, educational tools, or customer services, creating experiences that go far beyond traditional voice assistants. GPT-realtime is not just a “talking text AI,” but a voice agent that listens, acts, and responds seamlessly.

This could fundamentally change the world of work. Every company will prefer to use a more cost-effective, always available, and always accessible AI that sounds just as human as real people for service support. That is the future!

Why it matters: GPT-realtime finally establishes language as an interface. It lowers the barriers to use and opens up new markets where real-time AI will be indispensable.

Sources:

🔗 https://openai.com/index/introducing-gpt-realtime/

🔗 https://dev.to/czmilo/openai-gpt-realtime-complete-guide-revolutionary-breakthrough-in-voice-ai-2025-20m4?utm_source=chatgpt.com

🔗 https://www.reddit.com/r/OpenAI/comments/1n2kwvh/todays_gptrealtime_release/?utm_source=chatgpt.com

Stop wasting time on repetitive explanations. Guidde’s AI creates stunning video guides in seconds—11x faster.

How it works: Click capture on the browser extension, and Guidde auto-generates step-by-step video guides with visuals, voiceover, and a call to action.

Best part? It’s 100% free.

👉 Download the extension now

“Evals are important, but the vibes are hard to fake.” Anthropic has consistently outperformed OpenAI in terms of vibe.

The great computing power race: In a world where computing resources are becoming a scarce commodity, an article by ‘Sifted’ emphasizes that the UK has broken out of “AI tenant mode” with its own stack. Control over infrastructure means power - those who own their own computing chain determine the pace and direction of AI development. In the long term: strategic autonomy, less dependence on US patents or Asian manufacturing capacities.

The US government is serving up over 90 measures that reconcile innovation with deregulation - AI is to become national infrastructure. Everything is geared towards speed and global dominance. Critics warn of social costs: job losses, fake news, innovation without diversity. The foreign policy thrust? Enforce US standards, intensify regulatory rivalries internationally.

Anthropic is bringing in high-ranking voices from the military and political circles to align AI technology with democratic security. Strategy: technological dominance and security under democratic values. At its core: building a technology that becomes a national security asset -democratic states as guardians of responsible AI.

Exploring how AI is reshaping politics, policy, or governance? We’re featuring sharp analysis and research on the intersection of AI and political systems in Superintelligence, read by 200k+ people.

If you have significant comments on AI and politics (unemployment, geopolitics, proposed legislation) and would like to have this discussion presented in the newsletter, get in touch via a brief summary to [email protected] with the subject line “Politics Submission”. We’ll reach out if your work is a fit for a future issue.

It looks like we're in for a hot September. Qwen is preparing several releases!

Something is happening at OpenAI…