Physical Address
17033 Neubrandenburg
Germany
ChatGPT remembers what you did on the PC.
With Computer History ChatGPT can now access your past desktop activity — for even more personalized responses. Open Source
OpenAI has introduced one of the most significant extensions for ChatGPT: Computer History. The AI can now look back at what you have done on your computer — and incorporate this context into conversations.
This means ChatGPT no longer just understands your current messages, but also the broader context of your work. Browsing history, open documents, previous sessions—everything is factored into the responses. A massive step towards truly personalized AI assistants.
What that means
- Deeper context for each request
- No more repeated explanations
- Seamless continuation of previous work
- Genuine memory effect across sessions
Qwen 3.8 is now running locally
🖥️ Locally executable
No cloud required — runs directly on your hardware
⚡ High performance
Qwen 3.8B offers strong reasoning capabilities.
Alibaba's Qwen 3.8 It can now be run entirely on your own hardware. The model delivers impressive performance without cloud dependency — ideal for data-sensitive environments and developers who want full control over their AI infrastructure.
Manus is free until August 25th.
The AI agent Manus Opens its platform completely — without a paywall. Try it now while the offer lasts.
Free access runs until August 25, 2025. Tool tip
Manus, the ambitious AI agent from China, has made a sensational decision: Until the August 25 The entire service is completely free of charge.
This is a tremendous opportunity for anyone who wants to try out one of the most powerful autonomous AI agents on the market risk-free. Manus can independently plan, execute, and complete tasks—from web research to code generation.
Try it now while the offer lasts — such an opportunity rarely opens.
Interview → finished text in minutes
WhisperAI It automatically transforms hours-long interviews into structured texts with tasks — a real time saver for content teams.
Transforming a one-hour interview into a structured text with clear tasks—in just a few minutes. That's exactly what it does. WhisperAI.
Audio transcription
Upload, sit back — the service transcribes hours of recordings with impressive accuracy.
Finished text
A readable, structured text is automatically generated from the raw transcript — no manual editing is required.
Task extraction
WhisperAI automatically identifies tasks and action items from the conversation and presents them clearly.
More information at whisperai.com
Suno Studio 2 is here
The popular AI music generator Suno has fundamentally revamped his studio — more control, better quality, new production tools.
Suno, the leading AI music generator, has Studio 2 Released. The new version brings significantly improved sound quality, more precise control over style and structure, and professional-sounding productions — directly from the text description.
Studio 2 This is a clear statement: AI music is leaving the demo phase and becoming a serious creative tool.
✨ Gemini 3.7 Flash
Google releases the next update just 3 weeks after Gemini 3.6 Flash.
Google has Gemini 3.7 Flash published — only three weeks after Gemini 3.6 Flash. This update pace is unusual even for Google and shows how intense the competition in the Flash segment has become.
Faster iteration cycles mean that developers' models improve incredibly quickly. Regular testing is essential.
ChatGPT × Higgsfield
Higgsfield in ChatGPT
Video AI directly in chat — Higgsfield is officially integrated into ChatGPT
Higgsfield Layers
New feature: Split generated videos into editable layers
Higgsfield → Video prompt → Video in chat
With the official integration of Higgsfield in ChatGPT The need to switch between text and video AI is eliminated. Users can generate videos directly in the chat window—without switching platforms or requiring API knowledge. This is a significant step towards... All-in-one AI platform.
Agents take over the browser
LTX 2.5 — Open source video in native 4K HDR
LTX Video 2.5 generated 4K HDR with multi-shot scenes — and runs on its own hardware. The most impressive open-source video model to date.
LTX Video 2.5 is the new open video generation model with native 4K HDR, Multi-scene support and the ability to run it on custom hardware. A milestone for open-source video AI.
Grok 4.6 is catching up
Grok 4.6 achieves the same score as GPT-5.6 Sol in the Artificial Analysis General Index. The AI race is getting closer.
Grok 4.6
xAI — Equalities reached
GPT-5.6 Sol
OpenAI — Previous leader
The great showdown
Grok 4.6 According to xAI, Artificial Analysis Index has caught up and is now on par with GPT-5.6 Sol. The benchmark duel between OpenAI and xAI is getting closer and closer — and for users, this means more choice with top quality.
The gap between the leading models is shrinking. Differentiation increasingly occurs through features, integration, and price—no longer through raw intelligence.
New models at a glance
Seedream 5.0 Pro
Seedream 5.0 Pro now masters the decomposition of generated images into separate, editable layers. What began at Higgsfield is becoming the industry standard: AI-generated images are no longer monolithic, but fully editable.
🧠 GLM-5.3 by Z.ai
Z.ai has GLM-5.3 They presented the next generation of their language model. The model is under Z.ai available and builds on the proven GLM architecture — with improved reasoning capabilities and stronger multilingual support.
🤖 DeepSeek V4 Pro
The most important update concerns agent work. DeepSeek promises a significant increase in quality in real-world work scenarios, where the model independently uses tools and executes action sequences.
Those were your 16 AI news items 🚀
From local models to browser agents — AI is evolving at breakneck speed. Share this post with someone who wants to stay up to date!
💬 Comment
Which news stories do you find most exciting?
🔁 Share
Tag someone who should read this!
