- Horizon AI
- Posts
- ChatGPT Can Now Remember Everything You Ever Told It š§
ChatGPT Can Now Remember Everything You Ever Told It š§
Turn hours of literature searches into seconds with AI-powered research context

Welcome to another edition of Horizon AI,
OpenAI is enhancing ChatGPTās memory as part of its goal to create āAI systems that get to know you over your life,ā furthering its push toward a more personalized assistant.
Letās get started!
Read Time: 4.5 min
Here's what's new today in the Horizon AI
ChatGPT Can Now Reference All Previous Chats as Memory
Mira Muratiās AI Startup Targets $2B Seed Round
AI Tutorial: Turn hours of literature searches into seconds with AI-powered research context
AI Tools to check out
The Latest in AI and Tech š”
AI Findings/Resources
AI News
OPENAI
ChatGPT Can Now Reference All Previous Chats as Memory

OpenAI has introduced an update to the memory feature in ChatGPT: it can now reference all of your past conversations, even those not manually saved, to better inform future responses.
Details:
ChatGPTās memory feature has been around for over a year. It could already remember things like your preferences or specific details you asked it to remember, carrying them over from one chat to the next.
With the new upgrade, it now automatically builds on everything it knows from your previous conversations ā allowing for smoother, more contextualized, and helpful interactions.
As with the previous memory feature, users can opt out. They can also use temporary chat to have a conversation that wonāt use or affect memory.
The update is currently rolling out to Pro users and will be available to Plus users soon. However, it wonāt be available in the EU, UK, Switzerland, Norway, Iceland, or Liechtenstein.
This is part of OpenAIās broader strategy to position ChatGPT not as a mere AI chatbot, but more like a personal assistant that can seamlessly integrate into users' lives.
THINKING MACHINES LABS
Mira Muratiās AI Startup Targets $2B Seed Round

Ex-OpenAI CTO Mira Muratiās new AI startup named Thinking Machines Lab is reportedly aiming to raise $2 billion in seed funding, potentially valuing it at over $10 billion.
Details:
Thinking Machines Lab recently emerged from stealth mode and, despite having no product or revenue yet, it seems to have attracted investors with its ambitious vision and dozens of high-profile AI researchers in its ranks.
According to Business Insider, the startup will double its target for the seed funding round, making it one of the largest in history.
Murati spent six years helping shape projects like ChatGPT before leaving OpenAI in September 2024. Just six months later, she launched her own AI startup and secured key hires, including senior OpenAI executives and talent from Google DeepMind and Character AI.
Thinking Machines Lab now has the team, the funding, and the ambition to make an impact in AI research with its goal of creating AI systems that are āmore widely understood, customizable, and generally capableā than those currently available.
AI Tutorial
Turn hours of literature searches into seconds with AI-powered research context

Go to the new AlphaXiv Assistant and sign in.
Type in your research question or topic, like: 'What are the latest breakthroughs in RL fine-tuning?'
Youāll get comprehensive lit reviews with trending papers automatically included.
AI Tools to check out
š§ Google Illuminate: Transform your content into engaging AIāgenerated audio discussions
š Dezbor: Easily create dashboards or admin panels with AI and without coding.
š„ļø Screen Studio: Create beautiful screen recordings in minutes.
ā Conveyor: Share security documents and answer customer questions instantly with 95%+ accurate AI.
š Hello8: Translate your videos into 29+ languages with the latest AI technology.
AI Findings/Resources
š How to build a self-service agent for your website with Copilot Studio
š MIT study finds that AI doesnāt, in fact, have values
š Energy demands from AI datacentres to quadruple by 2030, says report
š AI-generated images can exploit how your mind works ā hereās why they fool you and how to spot them
The latest in AI and Tech
āOpenAI is poised to launch GPT-4.1, the next iteration of its AI model, as early as next week. This update is expected to replace the current GPT-4o model and will be accompanied by smaller versions, GPT-4.1 mini and nano. ā
The new initiative is aimed at developing AI benchmarks that more accurately reflect real-world applications in sectors like law, finance, and healthcare.
The company acknowledges that existing benchmarks often fail to capture the complexities of practical use cases and can be prone to manipulation.
The company is starting to roll out an update to its Creator Music marketplace that lets creators generate original tunes using AI prompts.
āSafe Superintelligence (SSI) has partnered with Google Cloud to utilize its TPU chips to āaccelerate its research and development efforts toward building a safe, superintelligent AI.ā
Thatās a wrap!
We'd love to hear your thoughts on today's email!Your feedback helps us improve our content |
Not subscribed yet? Sign up here and send it to a colleague or friend!
See you in our next edition!
Gina š©š»āš»