šŸŽ“ OpenAI reportedly plans to unveil "Ph.D.-level super-agents"

Also: Google's new 'Titans' neural architectures

Welcome, AI enthusiasts

Breaking news from Axios indicates that OpenAI is gearing up for a big launch ā€” a PhD-level agent. Oscar-nominated writer-director Paul Schrader posed some script ideas to AI chatbot ChatGPT and he is now confident that it can write ā€œoriginalā€ screenplays. Lastly, Google's Titans is a new family of neural architectures designed to address the limitations of existing models (mainly Transformers). Letā€™s dive in!

In todayā€™s insights:

  • OpenAI reportedly plans to unveil "Ph.D.-level super-agents" at the end of January

  • Googleā€™s new AI architecture ā€˜Titansā€™ can remember long-term data

  • Paul Schrader claims ChatGPT writes "original" and "fleshed out" movie scripts faster than writers

Read time: 4 minutes

šŸ—žļø LATEST DEVELOPMENTS

Source: Midjourney prompted by Evolving AI

Evolving AI: OpenAI plans to showcase advanced AI agents capable of solving PhD-level problems at the end of January, according to Axios.

Key Points:

  • OpenAIā€™s Sam Altman will meet U.S. officials on Jan. 30 to discuss new AI advances.

  • The presentation will feature ā€œPhD-level super-agentsā€ designed to handle complex tasks.

  • Altman will also address AIā€™s role in economic growth and computing challenges.

Details:

OpenAI is going to show off super-smart AI helpers that can do tough jobs really well. These AI systems can handle big chunks of information, think through choices, and work toward goals. On January 30, Sam Altman will present these new tools in Washington and talk about OpenAIā€™s plan for AI and the economy. People in the industry say AI is improving faster than expected. According to the report, some OpenAI workers have told their friends theyā€™re both ā€œjazzed and spookedā€ by the recent progress. Some question if these tools are reliable, but OpenAI wants to prove they work. In a recent blog, Altman wrote about entering an ā€œIntelligence Age,ā€ where powerful systems will make AI available to more people while reducing risks.

Why It Matters:

If this breakthrough is real, itā€™s far more than an advancement in AI ā€” itā€™s a societal shift. This has the potential to redefine how knowledge is used and tasks are automated, impacting education, work, and governance. The bigger question: is anyone prepared for this change?

TOGETHER WITH ARTISAN

Never Miss Another Warm Lead With Our AI BDR

Never miss a hot lead again. Our AI BDR Ava tracks intent signals across the webā€”triggering perfectly timed outreach when prospects are ready to buy.

She operates within the Artisan platform, which consolidates every tool you need for outbound:

  • 300M+ High-Quality B2B Prospects, including E-Commerce and Local Business Leads

  • Automated Lead Enrichment With 10+ Data Sources

  • Full Email Deliverability Management

  • Multi-Channel Outreach Across Email & LinkedIn

  • Human-Level Personalization

Free up your sales team to focus on high-value interactions and closing deals, while Ava handles the time-consuming tasks.

Source: Mert Alper Dervis/Anadolu via Getty Images

Evolving AI: Paul Schrader, writer of the legendary movie Taxi Driver, claims ChatGPT can quickly generate detailed and original movie scripts.

Key Points:

  • Paul Schrader asked ChatGPT to come up with movie ideas for famous directors, including himself.

  • The results amazed him, showing how AI can create creative and detailed ideas in seconds.

  • Schraderā€™s comments add to the ongoing discussion about AIā€™s role in making movies.

Details:

Oscar-nominated filmmaker Paul Schrader tested ChatGPTā€™s ability to create movie ideas. He gave it prompts for himself and famous directors like Quentin Tarantino and Martin Scorsese. Schrader said the chatbot came up with creative and complete concepts in seconds. His comments on social media started a debate, with some saying AI depends on existing work. Schrader sees AI as a tool to make the creative process faster, possibly saving months of work. This isnā€™t the first time Schrader has spoken positively about AI in filmmaking. During the WGA strike in 2023, he said AI could write a CSI episode quickly by using years of show history.

Why It Matters:

Schraderā€™s last sentence is probably the most telling of them all: ā€œWhy should writers sit around for months searching for a good idea when AI can provide one in seconds?ā€ It raises a key question about the role of human creativity in the age of AI. Will AI become a trusted partner in the creative process, or will it overshadow the human touch that defines art?

Evolving AI: Google introduces Titans, a neural network that extends memory without skyrocketing costs.

Key Points:

  • Titans combines short-term and neural long-term memory for efficient learning.

  • It handles millions of tokens while cutting memory and compute costs.

  • Titans outperforms larger models in long-sequence tasks.

Details:

The memory in Titans works by tracking two types of ā€œsurpriseā€ ā€” immediate surprise (how unexpected is the current word?) and accumulated surprise (what patterns of unexpected things have been appearing?). This helps the system decide what information to keep and what to forget. What makes this clever is that itā€™s set up as a gradient descent problem, allowing efficient parallel processing even though itā€™s naturally sequential. What stands out is how the memory connects to the main model. Among three tested methods, the most effective was using memory as extra context alongside the input. This allows the attention mechanism to decide whether to use memory or immediate context. Because memory tokens are included during both training and inference (the process of making predictions or decisions using a trained model), the memory and main model learn together, even as the memory model adapts dynamically during inference.

Why It Matters:

This design lets Titans handle sequences over 2 million tokens long while outperforming traditional transformers. It even matches GPT-4 in some long-context reasoning tasks, using far fewer parameters. Could this be the end of Transformer based LLMs?

 šŸ‘€ Click on the image you think is real

šŸŽ­ Oscar hopeful The Brutalist comes under fire for using AI.

šŸŒ Explained: Generative AIā€™s environmental impact.

šŸŽ® AI startup Character AI tests games on the web.

āš”ļø The Pentagon says AI is speeding up its ā€˜kill chainā€™.

šŸ—ŗļø Google Maps is turning 20 ā€” itā€™s mapping three more countries and adding AI capabilities.

šŸŽ“ ā€˜A death penaltyā€™: Ph.D. student says U of M expelled him over unfair AI allegation.

šŸ“° Google signs deal with AP to deliver up-to-date news through its Gemini AI chatbot.

šŸ“ˆ Trending AI Tools

  • šŸ’»ļø Sana - Your ultimate work companion. It's the powerhouse assistant you need for everything from document analysis to task automation (link)

  • āœļø ProDream - An academic writing assistant with outlining, editing, citation formatting, proofreading, and plagiarism checking (link)

  • šŸ“‹ Jamie - A tool to automate meeting note-taking by providing summaries, transcripts, and action items from audio recordings (link)

  • šŸ’¬ Docus - A chat assistant for personalized health consultations and second opinions (link)

  • šŸŒ± Empaithy - An app for emotional support and self-reflection through micro-journaling (link)

Reply

or to participate.