Innovation on Carles Abarca

AI Agents Are No Longer Assisting Scientists. They Are Doing the Science.

Sun, 15 Mar 2026 00:00:00 +0000

In March 2026, something shifted. Not a bigger model. Not a higher benchmark. Something deeper: AI agents stopped being tools that help scientists and started producing scientific knowledge on their own.

This week, three events converged, and I believe they mark a tipping point with no return. The ShipSquad team documented it brilliantly in their analysis “AutoResearch, OpenClaw, Claude Opus 4.6: AI Agents Are Now Doing the Science”, and it inspired me to dig deeper into what this means for scientific research as we know it.

The facts
#

Andrej Karpathy released AutoResearch — a 630-line Python framework that lets an AI agent run hundreds of machine learning experiments autonomously on a single GPU. You let it run overnight. You wake up to a better model and a complete discovery log. In 48 hours, the agent found ~20 improvements no human had identified, cutting GPT-2 training time by 11%.

Anthropic’s Claude Opus 4.6 discovered 22 zero-day vulnerabilities in Firefox — 14 of them high-severity — in just two weeks. For context: those 14 represent nearly one-fifth of all serious Firefox vulnerabilities patched throughout 2025. An AI model matched months of human security research.

Sakana AI unveiled The AI Scientist v2 — an agentic system that generates hypotheses, designs experiments, executes them, analyzes results, and writes the complete scientific paper. The result: the first entirely AI-generated scientific paper accepted through peer review at an academic workshop.

Three events. The same week. The same conclusion: AI agents are now producing original knowledge.

AI has discovered before, but never on its own
#

To understand why this moment is different, trace the evolution:

Era 1: AI as microscope (2018-2023)
#

AI amplified human researchers’ capabilities. Humans asked the questions, designed the experiments, and AI processed data at impossible scales.

AlphaFold (2020-2024) predicted the structure of all 200 million known proteins. A problem that had resisted solution for 50 years. Hassabis and Jumper won the 2024 Nobel Prize in Chemistry for it. But the question — “can we predict protein structures?” — was formulated by humans.
Halicin (2020) — MIT researchers used AI to screen 100 million chemical compounds and discovered a new antibiotic capable of killing multi-drug-resistant bacteria, including the dreaded Acinetobacter baumannii. But the experimental design was human.
GNoME by DeepMind (2023) discovered 2.2 million new crystals, including 380,000 stable materials — multiplying by 10x everything humanity had found in the entire history of materials science. But the evaluation framework was designed by Google researchers.

Era 2: AI as colleague (2024-2025)
#

AI began proposing hypotheses and designing experiments, but under human supervision.

FunSearch by DeepMind (2024) used an LLM to discover new solutions to open problems in pure mathematics — the first time a language model made a genuine discovery in formal sciences.
Insilico Medicine got its molecule rentosertib — designed entirely by generative AI — through a Phase IIa clinical trial with positive results for idiopathic pulmonary fibrosis. From concept to human testing in under 30 months, when the traditional process takes 10-15 years.
MIT Antibiotics-AI Project (2025) moved from discovering existing antibiotics to designing entirely new molecules using generative AI capable of killing resistant bacteria. No longer searching for needles in a haystack; they are manufacturing new needles.

Era 3: AI as autonomous researcher (2026 →)
#

And now, in March 2026, we cross the threshold: AI formulates its own questions, designs its own experiments, executes them, and produces papers accepted by human reviewers. Not science fiction. It is AutoResearch, AI Scientist v2, and Claude Opus doing independent security research.

The Great Acceleration: when AI researches faster than humans can read
#

I considered making a safer prediction — forecasting the percentage growth of AI-authored scientific papers. Instead, here is the unfiltered version.

Before 2030, AI agents will have produced more verifiable scientific discoveries in materials science, drug discovery, and computational mathematics than all human researchers combined in those same disciplines.

This is not hyperbole. It is arithmetic.

Consider the numbers: GNoME discovered in weeks 380,000 stable materials that all of humanity took centuries to accumulate (barely 48,000 by 2023). AutoResearch runs 100 experiments per night — months of a PhD student’s work. And AI Scientist v2 can generate a complete scientific paper in hours, not months.

Now scale that. Not one agent, but thousands. Not one night, but every night. Not one domain, but all of them.

The Compound Acceleration phenomenon
#

What we are witnessing is not linear acceleration. It is compound acceleration: each agent discovery feeds the next research cycle. One agent discovers a new material → another simulates its properties → another designs applications → another writes the paper. In parallel. 24/7. No vacations, no ego, no departmental politics.

Human science operates at publication speed: one paper every 6-18 months. Agentic science will operate at computation speed: one discovery per minute.

Problems we will solve decades ahead of schedule
#

This acceleration is not just quantitative. There are problems we believed were decades away that research agents could solve much sooner:

Antibiotic resistance — AI is already designing new molecules against superbugs. With autonomous agents iterating 24/7 on thousands of variants, we could have a complete new generation of antibiotics before 2030. The WHO estimated that by then, superbugs would kill 10 million people per year. Agents could get there first — for the benefit of all humanity.
Nuclear fusion — The greatest challenge is controlling plasma. DeepMind already used AI to control plasma shape in the TCV tokamak. Autonomous agents simulating and optimizing millions of magnetic configurations could compress decades of experimental research into years.
Cellular aging — AlphaFold solved protein structures. The next frontier is understanding the interactions between proteins, genes, and cellular processes that cause aging. It is a problem of massive combinatorial complexity — exactly the kind of problem where agents excel.
New energy materials — GNoME already opened the door with 380,000 stable materials. Agents systematically exploring that space could find the room-temperature superconductor, the perfect battery electrolyte, or the catalyst that makes industrial carbon capture viable.

The question is no longer whether AI will surpass humans in scientific output. It is when. And my answer is: in many fields, it is already happening. We just have not updated our metrics to measure it.

The consequences nobody wants to discuss
#

1. The end of “batch research”
#

Today, a researcher can run perhaps 5-10 experiments per week. AutoResearch demonstrates that an agent runs 100 per night. When research shifts from a sequential human process to a continuous autonomous one, knowledge production will multiply by orders of magnitude.

2. The radical democratization of discovery
#

Karpathy proved that a single GPU and 630 lines of code are enough for autonomous research. A PhD student in Monterrey, Lagos, or Bangalore can now compete in knowledge production with a Stanford lab. The barrier is no longer budget; it is the ability to formulate good questions and direct agents with precision.

3. The “Orchestra Conductor” as the new scientific role
#

The scientist of the future will not be the one pipetting in the lab or writing code. It will be the one who “programs in Markdown” — who writes the precise instructions that guide squadrons of autonomous agents. This is exactly what Karpathy demonstrates with his program.md file: the future of directing research is not writing better code, but writing better programs in natural language.

This paradigm shift is not new. I explored it in my article “The Next AI Wave: Agents” — where I argued that autonomous agents represent a fundamental shift from assistive AI to autonomous execution. The same applies to science: the researcher becomes the conductor of research agent orchestras.

4. The peer review crisis
#

If AI Scientist v2 already generates papers accepted at workshops, how long until it produces papers accepted at top conferences? And how will we distinguish between a paper written by an agent and one written by a human? The peer review system, designed to evaluate human work, is unprepared for a world where papers are generated at industrial speed. We will need new evaluation frameworks — and possibly, AI agents reviewing the papers of other AI agents.

What this means for universities
#

The institutions that most quickly integrate autonomous research agents into their labs will lead knowledge production in the next decade. It is not about buying bigger GPUs. It is about:

Training “research agent directors” — scientists who know how to formulate questions and direct AI squadrons.
Building autonomous experimentation infrastructure — labs where agents can run thousands of experiments without continuous human oversight.
Redefining authorship and intellectual property — if an AI agent discovers a molecule that cures a disease, who owns the discovery?
Measuring scientific output differently — current indicators (papers, citations, h-index) are metrics designed for human speed. We need metrics that capture agentic velocity.

The race has started. And the window of opportunity to position yourself is now.

Conclusion
#

In just six years, AI went from predicting protein structures to discovering antibiotics, from designing materials to writing scientific papers, from finding known vulnerabilities to discovering zero-days. The trajectory is clear and accelerating exponentially.

AI agents will not replace scientists. They will make scientists who do not use them irrelevant. And that transition, unlike the one we lived through in the software industry, will not take a decade. It will take months.

Welcome to the era of the Great Scientific Acceleration. Those who understand it first will lead the science of the future. Those who ignore it will read about it in papers written by machines.

🤖 Go deeper with AI
#

Want to explore further? Click your favorite AI with a ready-made prompt:

On the scientific acceleration with AI agents: ChatGPT · Perplexity · Claude

On the future of peer review: ChatGPT · Perplexity · Claude

On the impact for universities: ChatGPT · Perplexity · Claude

Neuromorphic Computing: The Future of Artificial Intelligence

Wed, 14 Aug 2024 00:00:00 +0000

In the midst of the AI revolution, neuromorphic computing fuels the possibility of reaching singularity, or general AI. Inspired by the human brain, this technology promises to revolutionize our conception of what we have so far considered intelligent life.

What is Neuromorphic Computing?
#

It is an approach that emulates the brain’s neural networks to improve the efficiency and adaptability of traditional computing technologies based on the Von Neumann architecture. Chips like IBM’s TrueNorth and Intel’s Loihi exemplify this technology, allowing parallel task processing, consuming less energy, and adapting to new tasks, much like how the human brain learns.

Key Applications
#

AI and ML: Enhancements in image and voice recognition, and predictive analysis.
Robotics: More efficient and natural interactions, almost “human” robots.
IoT: Smart devices that autonomously respond in real-time.

A Future of Unlimited Possibilities
#

From enhanced AI capabilities to future integration with quantum computing, the potential of neuromorphic computing is immense. IBM, Intel, and leading universities are at the forefront of this development, paving the way toward technology that will offer nearly unlimited possibilities and raise unprecedented ethical challenges.

Conclusion
#

Neuromorphic computing not only redefines what is possible in technology but also aligns computational capabilities with human cognitive processes, opening the doors to an exciting and uncertain future.

What to Expect from Multimodal AI in 2024 and 2025

Wed, 05 Jun 2024 00:00:00 +0000

The future of AI is incredibly exciting, and 2024 is set to bring some amazing advancements into our everyday lives. Multimodal AI agents, which can understand and process text, images, audio, and video all at once, are going to change how we interact with technology in profound ways.

Seamless Communication
#

Imagine having a virtual assistant that doesn’t just respond to your voice commands but also understands your gestures and facial expressions. Whether you’re cooking, working out, or just relaxing at home, these AI agents will make interacting with your devices more intuitive and natural.

Smarter Home Assistants
#

Your home assistant will become a true member of the family. It will recognize when you’re feeling down and play your favorite music, suggest a movie based on your recent viewing habits, or even help you troubleshoot a problem by visually guiding you through the steps.

Enhanced Shopping Experiences
#

Shopping online will be more personalized and engaging. These AI agents can help you find clothes that match your style, fit your body shape, and even suggest outfits based on your existing wardrobe. They can also provide real-time support during your shopping experience, making it feel like you have a personal shopper at your side.

Health and Wellness
#

From virtual fitness trainers that can correct your form through video analysis to mental health apps that understand your mood through voice and text, multimodal AI will support your well-being in more interactive and personalized ways.

Learning and Education
#

Education will become more accessible and tailored to individual needs. Whether it’s helping kids with homework through interactive video sessions or enabling adults to learn new skills with personalized, multimedia lessons, these AI agents will make learning more effective and enjoyable.

Entertainment and Creativity
#

Multimodal AI will transform how we create and consume entertainment. Imagine AI that can help you compose music by understanding your mood and preferences, or create visual art based on your descriptions and sketches. Your favorite shows and games will become even more immersive, adapting to your reactions and feedback in real-time.

Final Thoughts
#

As we approach 2025, the integration of multimodal AI into our daily lives promises to make technology more accessible, personal, and helpful than ever before. Whether at home, at work, or at play, these advancements will enhance our experiences and open up new possibilities.

Goodbye Prompt Engineering

Tue, 30 Apr 2024 00:00:00 +0000

As we come to understand the possibilities offered by the wave of AI-related technological advances, new ways of relating to these technologies are emerging. Our interactions with generative artificial intelligence are transforming, marking the end of the “prompt engineering” era – a short-lived “era” in which some claimed that mastering the art of “prompting” would be key to benefiting from this new technology. I never subscribed to the notion that “prompt engineers” would lead AI adoption, and here are some reasons to anticipate the gradual disappearance of the ephemeral “art of prompt engineering”:

From Interaction to Collaboration
#

The traditional model of interacting with large language models through specific commands is evolving. The future points toward interfaces with autonomous, goal-directed agents. This shift promises greater alignment with human intentions and a significant increase in AI decision-making autonomy.

Enhanced Collaboration
#

These advanced AI agents are designed to understand and anticipate needs, making them perfect partners in diverse professional environments. Whether optimizing decision processes or offering predictive insights, AI is moving toward a more proactive role, eliminating the need for detailed and repetitive command engineering.

Real-World Applications
#

Imagine an AI that not only responds to commands but also initiates actions aligned with established goals in fields like customer service, healthcare, and finance. This proactive capability could redefine efficiency and effectiveness across various industries. We will progressively see specialized autonomous agents for specific tasks rather than interacting with general-purpose models.

Innovation on Carles Abarca

AI Agents Are No Longer Assisting Scientists. They Are Doing the Science.

The facts #

AI has discovered before, but never on its own #

Era 1: AI as microscope (2018-2023) #

Era 2: AI as colleague (2024-2025) #

Era 3: AI as autonomous researcher (2026 →) #

The Great Acceleration: when AI researches faster than humans can read #

The Compound Acceleration phenomenon #

Problems we will solve decades ahead of schedule #

The consequences nobody wants to discuss #

1. The end of “batch research” #

2. The radical democratization of discovery #

3. The “Orchestra Conductor” as the new scientific role #

4. The peer review crisis #

What this means for universities #

Conclusion #

🤖 Go deeper with AI #

Neuromorphic Computing: The Future of Artificial Intelligence

What is Neuromorphic Computing? #

Key Applications #

A Future of Unlimited Possibilities #

Conclusion #

What to Expect from Multimodal AI in 2024 and 2025

Seamless Communication #

Smarter Home Assistants #

Enhanced Shopping Experiences #

Health and Wellness #

Learning and Education #

Entertainment and Creativity #

Final Thoughts #

Goodbye Prompt Engineering

From Interaction to Collaboration #

Enhanced Collaboration #

Real-World Applications #

The facts
#

AI has discovered before, but never on its own
#

Era 1: AI as microscope (2018-2023)
#

Era 2: AI as colleague (2024-2025)
#

Era 3: AI as autonomous researcher (2026 →)
#

The Great Acceleration: when AI researches faster than humans can read
#

The Compound Acceleration phenomenon
#

Problems we will solve decades ahead of schedule
#

The consequences nobody wants to discuss
#

1. The end of “batch research”
#

2. The radical democratization of discovery
#

3. The “Orchestra Conductor” as the new scientific role
#

4. The peer review crisis
#

What this means for universities
#

Conclusion
#

🤖 Go deeper with AI
#

What is Neuromorphic Computing?
#

Key Applications
#

A Future of Unlimited Possibilities
#

Conclusion
#

Seamless Communication
#

Smarter Home Assistants
#

Enhanced Shopping Experiences
#

Health and Wellness
#

Learning and Education
#

Entertainment and Creativity
#

Final Thoughts
#

From Interaction to Collaboration
#

Enhanced Collaboration
#

Real-World Applications
#