Updates and Insights for Today

1. AlphaEvolve: The AI That Outcodes Humans

2. GPT-4.1 Lands in ChatGPT: Faster, Smarter & Ready to Code

3. Why AI Still Fails at “Not”

4. The latest in AI tech

5. AI Tutorials: An AI Agent That Builds Other AI Agents? Yes, Really.

6. AI tools to checkout

AI News

AlphaEvolve: The AI That Outcodes Humans

Quick Summary:

DeepMind has launched AlphaEvolve, an AI coding agent powered by Gemini models. It designs advanced algorithms, surpassing human-devised solutions in areas like matrix multiplication and data center optimization. This marks a significant step in AI’s role in algorithm development.

Key Insights:

• AlphaEvolve combines Gemini’s language models with automated evaluators to discover and refine algorithms.

• It outperformed the 56-year-old Strassen algorithm in matrix computations.

• The AI has optimized tasks in data center scheduling, chip design, and language model efficiency.

• Its evolutionary approach ensures the generation of novel and provably correct solutions.

Why It’s Relevant:

AlphaEvolve demonstrates AI’s potential to innovate beyond human capabilities in algorithm design. By automating the creation of efficient algorithms, it can revolutionize computing tasks, leading to faster and more efficient systems. This advancement could have widespread implications across various industries.

📌 Read More: DeepMind

Our Partner Today

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

• Get daily AI news, tools, and tutorials

• Learn new AI skills you can use at work in 3 mins a day

• Become 10X more productive

Sign up and start mastering AI

GPT-4.1 Lands in ChatGPT: Faster, Smarter & Ready to Code

Quick Summary:

OpenAI has officially released GPT-4.1, now integrated into ChatGPT for Plus, Pro, and Team users. This new model excels in coding tasks, offering enhanced speed and precision, and introduces a ‘mini’ version for quicker interactions.

Key Insights:

• GPT-4.1 and its ‘mini’ variant are optimized for coding and instruction-following tasks.

• The models support a context window of up to 1 million tokens, significantly surpassing previous versions.

• GPT-4.1 is approximately 40% faster and uses 80% fewer resources per request compared to GPT-4o.

• The release includes a shopping feature within ChatGPT, allowing product selection and purchase directly through the interface.

Why It’s Relevant:

The integration of GPT-4.1 into ChatGPT marks a significant advancement in AI-assisted coding and user interaction. With its enhanced capabilities and efficiency, GPT-4.1 is poised to streamline workflows for developers and professionals alike. The addition of e-commerce functionality within ChatGPT also indicates OpenAI’s move towards more interactive and multifunctional AI applications.

📌 Read More: Wired

Why AI Still Fails at “Not”

Quick Summary:

A new study finds that even top-performing vision-language models like GPT-4V and Claude struggle to understand simple negative commands. These flaws show a major gap in how today’s AI interprets basic logic — and it could be a serious problem.

Key Insights:

• AI models fail to correctly answer prompts with negation like “Which animal is not a cat?”

• Even state-of-the-art systems often ignore negation words such as not, without, or except.

• The study used over 800 controlled prompts and found error rates as high as 80%.

• This issue impacts model reliability in safety-critical areas like medical and legal AI.

Why It’s Relevant:

Negation is a core part of human language — and the fact that AI models miss it so often is alarming. These blind spots can lead to serious misinterpretations. This study sheds light on a key weakness in multi-modal AI and raises the bar for future model design. Until this is fixed, AI will keep making very human-unfriendly mistakes.

📌 Read More: MIT

 AI Tutorials

I built an AI Agent that BUILDS AI Agents in N8N (Easy!)

Quick Summary:

A new workflow tool lets users auto-generate fully working AI agents by just answering a few simple questions. It builds structured prompts, selects tools, and creates real automation flows — no coding required.

Key Insights:

• You can build custom AI agents via drag-and-drop without needing dev skills.

• The system auto-generates prompts using a framework of goals, tools, and outputs.

• It selects APIs and builds workflows using tools like n8n and Google integrations.

• While it’s not perfect, it handles 90% of the setup automatically.

What Can I Learn?

• How to create working AI agents without writing code

• What inputs help the system build better prompts

• How tools are chosen and connected automatically

• What current limitations to expect (and how to fix them)

Which Benefits Do I Get?

• Saves hours of manual planning and setup

• Gives you ready-to-test automation flows

• Makes AI tools more accessible to non-developers

• Helps structure professional prompts for better AI performance

Why It’s Relevant:

Building useful AI agents used to take days and technical know-how. Now, one tool can do most of it for you — in minutes. This shift lowers the barrier for automation, helping solo founders, small teams, and creators turn ideas into working AI workflows. It’s a big leap toward everyday, customizable AI.

Here is the full Video Tutorial 👉 Click Here

The latest in AI tech

Stability AI & Arm Bring Audio AI to Your Pocket

Stability AI and Arm have released Stable Audio Open Small, a 341M parameter text-to-audio model optimized for on-device generation. It can produce up to 11 seconds of audio on a smartphone in less than 8 seconds. This open-source model runs entirely on Arm CPUs, making generative audio creation accessible on mobile devices. 

📌 Read More: Stability

Meta’s UMA Model Accelerates Scientific Research

Meta has unveiled the Universal Frontier Model for Atoms (UMA), an AI model trained on a massive dataset requiring 6 billion compute hours. UMA can perform quantum mechanical simulations 10,000 times faster than traditional methods, potentially revolutionizing drug and material discovery. 

📌 Read More: Semafor

Anthropic’s Claude Models to Enhance Reasoning

Anthropic is preparing to release new versions of its AI models, Claude Sonnet and Claude Opus, designed to improve complex task handling with less human input. These models can switch between reasoning and tool use, returning to reasoning mode to self-correct when necessary. 

📌 Read More: The Information

ChatGPT’s Role in Academia Sparks Debate

The integration of ChatGPT into college classrooms is causing controversy. Some students express concern over professors using AI-generated materials, leading to discussions about the ethical implications and the future role of AI in education. 

📌 Read More: NY Times

 AI Tools to check out

1. TheInfluencerAI: Pick traits or upload photos, and get your own photorealistic, consistent AI influencer.

2. CodeRabbit: Cut Code Review Time & Bugs in Half. Instantly.

3. Amie: Turn Notes into Workflows & much more

4. Scottie: Build Any AI Agent in 5 Minutes—No Code

Our Second Partner Today

Fact-based news without bias awaits. Make 1440 your choice today.

Overwhelmed by biased news? Cut through the clutter and get straight facts with your daily 1440 digest. From politics to sports, join millions who start their day informed.

Sign up now!

Thanks for sticking with us to the end!

We'd love to hear your thoughts on today's email!

Your feedback helps us improve our content

⭐⭐⭐Superb
⭐⭐Not bad
⭐ Could've been better

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Tom