- The Current ⚡️
- Posts
- ChatGPT Update Expands Memory — A LOT
ChatGPT Update Expands Memory — A LOT
Also, Canva revamps it’s whole design integrating AI into every corner

⚡️ Headlines
🤖 AI
Microsoft launches Recall preview for Windows Insiders - Microsoft's Recall feature, which captures screen snapshots to help users retrieve past activities, is now available to Windows Insiders in the Release Preview channel, signaling an imminent broader release for Copilot Plus PCs. [The Verge]
Mira Murati's AI startup seeks record $2B seed funding - Former OpenAI CTO Mira Murati aims to raise over $2 billion for her new venture, potentially marking the largest seed round in history, with a team comprising several ex-OpenAI colleagues. [TechCrunch]
OpenAI is winding down its GPT-4 model in ChatGPT - OpenAI is beginning to sunset GPT-4 in ChatGPT as it transitions users toward newer models with improved performance and functionality. [TechCrunch]
Inside the LLM: AI2’s OLMoTrace will trace the source - AI2 has launched OLMoTrace, a tool designed to make large language models more transparent by tracking how training data influences model outputs. [VentureBeat]
Fintech founder charged with fraud after AI shopping app revealed to be run by humans - A fintech CEO has been charged with fraud after it was discovered that their supposedly AI-powered shopping assistant was actually operated manually by workers in the Philippines. [TechCrunch]
Google DeepMind may merge Gemini and Veo AI models, says CEO - DeepMind’s Demis Hassabis reveals plans to potentially combine the Gemini and Veo AI models for a more unified multimodal system. [TechCrunch]
Writer launches AI-HQ platform betting on autonomous agents - Writer unveils AI-HQ, a new enterprise platform powered by AI agents designed to handle a wide range of internal business tasks autonomously. [VentureBeat]
NTT unveils ‘Physics of AI’ group and AI inference chip for 4K video - Japanese tech giant NTT has launched a new AI-focused research group and introduced a chip optimized for 4K video inference in AI applications. [VentureBeat]
🦾 Emerging Tech
Killing the IRS DeFi rule is an industry win—but temporary - The repeal of a controversial IRS rule targeting DeFi platforms marks a short-term victory for the crypto sector, though regulatory uncertainty remains. [CoinDesk]
🤳 Social Media
Meta uses TikTok sell-off uncertainty to court creators - Meta is capitalizing on regulatory threats against TikTok to lure creators with incentives and stability promises across Instagram and Facebook. [Social Media Today]
YouTube backs 'No Fakes Act' to fight AI impersonation - YouTube is throwing support behind the No Fakes Act to curb unauthorized AI-generated replicas, bolstering creator protections. [YouTube Blog]
🔬 Research
OpenAI launches BrowseComp to benchmark web-savvy AIs - OpenAI introduces BrowseComp, a new benchmark to test how well AI agents can find obscure and hard-to-locate information online. [OpenAI]
Microsoft introduces Debug Gym to teach AIs how to debug like humans - Microsoft Research unveils Debug Gym, a new environment for training AI tools to diagnose and fix programming errors in human-like ways. [Microsoft Research Blog]
Stanford study explores AI bots as emotional companions - Researchers at Stanford explore how AI bots could serve as emotional companions and impact users’ mental health and social interactions. [Social Media Today]
⚖ Legal
Elon Musk wants to be AGI 'dictator,' OpenAI tells court - In a legal filing, OpenAI accuses Elon Musk of trying to dominate AGI development and discredit the organization for not aligning with his vision. [Ars Technica]
🎱 Random
Canva unveils its biggest redesign in 13 years - Canva debuts a major redesign with AI-powered tools and new capabilities aimed at expanding its platform beyond design into business productivity. [Fast Company]
🔌 Plug-Into-This
OpenAI has expanded ChatGPT’s memory capabilities, enabling the AI to autonomously recall and incorporate details from users' entire chat histories. This enhancement allows for more personalized and context-aware interactions, with the feature currently rolling out to ChatGPT Plus and Pro users globally, excluding certain regions.

ChatGPT now automatically retains information from past conversations, enhancing its ability to provide tailored responses based on user preferences and interests.
Users maintain control over this feature, with options to manage or disable memory settings at any time.
The memory upgrade is being introduced to individual Plus and Pro subscribers, with plans to extend it to Team, Enterprise, and Edu accounts in the coming weeks.
This development marks a significant step toward more personalized AI assistants that adapt over time to individual users.
The feature excludes users in the UK, Switzerland, Norway, Iceland, and the European Economic Area during the initial rollout phase.
Starting today, memory in ChatGPT can now reference all of your past chats to provide more personalized responses, drawing on your preferences and interests to make it even more helpful for writing, getting advice, learning, and beyond.
— OpenAI (@OpenAI)
5:06 PM • Apr 10, 2025
🧠 Continuity and personalization look like the targets now as ChatGPT looks to cement it’s position at the fore of the AI industry chatbots. Privacy concerns abound, but 90% of us use Gmail still so it’s doubtful anyone will care enough to leave the comfort and convenience of the chatbot that knows everything about you!
Canva has launched Visual Suite 2.0, a comprehensive update that integrates generative AI-powered tools, expanding its platform beyond design into coding and spreadsheets. This positions Canva as a unified workspace solution, aiming to streamline workflows for its 230 million users.

Canva Sheets introduces a visually oriented spreadsheet tool that combines data import capabilities with AI-generated insights and interactive charts.
Canva Code allows users to generate websites and interactive content through text prompts, catering to those with minimal coding experience.
A centralized AI chatbot has been introduced to facilitate access to Canva's creative tools via voice or text commands.
The updated Photo Editor now offers advanced features, including background modifications and AI-generated environments, paralleling professional-grade software.
These enhancements aim to unify design and productivity tools, enabling teams to collaboratively produce various content types within a single platform.
This is wild. Canva Create 2025 just happened.
The new Canva will completely change design work with bonus AI magic.
10 unreal updates you can't miss:
— Shushant Lakhyani (@shushant_l)
10:00 AM • Apr 11, 2025
💪 By quickly integrating AI-driven features into its platform, Canva is set to make a real challenge to traditional productivity suites, offering a more intuitive and visually focused alternative for creative and business users alike.
A Greenpeace analysis has revealed that emissions from AI chip manufacturing increased by 357% in 2024, highlighting the environmental impact of the rapidly growing AI industry.

The production of semiconductors for AI applications saw a more than fourfold increase in emissions, outpacing the growth in electricity usage.
This rise is attributed to the escalating demand for AI services, leading to increased chip production and associated energy consumption.
The environmental footprint of AI chip manufacturing raises concerns about the sustainability of current industry practices.
Greenpeace's findings call attention to the necessity for greener technologies and more efficient manufacturing processes in the AI sector.
The report suggests that without intervention, the environmental impact of AI chip production will continue to escalate alongside industry growth.
BREAKING: Our new report with @DigiEconomist finds that AI companies like Nvidia are threatening climate progress. Emissions from AI chipmaking increased by more than 350% in just one year.
[1/4]
— Greenpeace East Asia (@GreenpeaceEAsia)
12:14 AM • Apr 10, 2025
🌍 The rapid expansion of AI technologies is set to bring environmental challenges even more to the forefront as both a political and business issue. As we charge forward, the importance of developing sustainable solutions within the tech industry's supply chain should also be rising considerably.
🆕 Updates
Mistral Le Chat is rolling out libraries in beta.
- It is similar to ChatGPT, Claude projects, and Perplexity Spaces.
- It supports file uploads with the following extensions: PDF, DOCX, PPTX, CSV, TXT, MD, and XLSX.— Patel Meet (@MeetPatelTech)
5:21 AM • Apr 11, 2025
🚀 Coming soon from #GoogleCloudNext: Gemini 2.5 Flash on Vertex AI.
This workhorse model is optimized for low latency & reduced cost—the ideal engine for responsive virtual assistants and real-time summarization tools where efficiency at scale is key → goo.gle/3YkUWdy
— Google Cloud Tech (@GoogleCloudTech)
4:12 PM • Apr 9, 2025
Gemini now supports YouTube video as the input in Google AI Studio.
Gemini can actually understand YouTube videos and generate multilingual transcriptions, video descriptions with time stamps, identify sarcasm and so much more.
And it works across languages, so crazy.
— Shubham Saboo (@Saboo_Shubham_)
2:27 AM • Apr 10, 2025
📽️ Daily Demo
Great to see our @GoogleDeepMind @GoogleAI AI co-scientist innovations (multi-agent idea generation tournament) ship fast with @googlecloud Agentspace!
Incredible shipping momentum at @Google! 🚢
Agentspace -cloud.google.com/products/agent…
AI co-scientist - research.google/blog/accelerat…
— Vivek Natarajan (@vivnat)
8:11 PM • Apr 9, 2025
Veo 2 is more than just a video generation tool in @googlecloud’s #VertexAI. 🎥
Here’s a rundown of its features. 🧵
— Google DeepMind (@GoogleDeepMind)
12:25 PM • Apr 10, 2025
🗣️ Discourse
love the feedback! - to MCP it is!
— Sundar Pichai (@sundarpichai)
9:29 PM • Apr 9, 2025
We’re excited to announce two @Nature publications from Project AMIE (Articulate Medical Intelligence Explorer), a research AI system optimized for diagnostic reasoning and conversations 💬
Paper 1: goo.gle/4lpQ8xg
Paper 2: goo.gle/3G4DNPe— Google Health (@GoogleHealth)
4:56 PM • Apr 9, 2025
Vertex AI is now the only platform with generative media models for video, image, speech + music:
- Lyria
- Veo 2
- Chirp 3
- Imagen 3Build a production-ready asset starting from a text prompt to an image to a complete video asset with music & speech → goo.gle/4lp7uug
— Google Cloud Tech (@GoogleCloudTech)
4:39 PM • Apr 9, 2025