In partnership with
Your prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning — constraints, edge cases, examples, tone — and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Try Wispr Flow free — works on Mac, Windows, and iPhone.
Here is one of the coolest uses of AI you probably have not heard about. The same kind of technology that powers image generators like DALL-E has been turned loose on one of the hardest problems in science: predicting the Earth's climate. And it just did something remarkable. A new AI model projected 100 years of global climate patterns in 25 hours, a task that normally takes the world's most powerful supercomputers weeks or even months. It is faster, cheaper, and nearly as accurate as the gold standard. Here is what happened and why it could change how we understand the future of the planet.
What the AI Actually Did
The breakthrough comes from an unexpected place: borrowing the technology behind AI art generators and pointing it at the atmosphere. The algorithms behind generative AI tools like DALL-E, when combined with physics-based data, can be used to develop better ways to model the Earth's climate. Computer scientists in Seattle and San Diego used this combination to create a model capable of predicting climate patterns over 100 years 25 times faster than the state of the art. Inside Higher Ed
The model has a name and some genuinely impressive numbers behind it. The model, called Spherical DYffusion, can project 100 years of climate patterns in 25 hours, a simulation that would take weeks for other models. Existing state-of-the-art models need to run on supercomputers. This model can run on GPU clusters in a research lab. Inside Higher Ed
That last part is the genuinely revolutionary detail. Where it takes about six months to run a physics-based model, theirs was able to produce equally good results in about two weeks. Traditional climate models are so demanding that they require national-scale supercomputers and months of run time. This AI model produces comparable results on the kind of computer hardware a university research lab can actually afford. The team developed it at UC San Diego and the Allen Institute for AI, and they were direct about the significance. "Data-driven deep learning models are on the verge of transforming global weather and climate modeling," the researchers wrote. UnsplashInside Higher Ed
The full UC San Diego announcement: https://today.ucsd.edu/story/accelerating-climate-modeling-with-generative-ai
Phys.org's breakdown of how it works: https://phys.org/news/2024-12-climate-combines-generative-ai-physics.html
Why This Is Genuinely a Big Deal
To understand why scientists are excited, you need to know how painful climate modeling has been. Climate simulations are currently very expensive to generate because of their complexity. As a result, scientists and policymakers can only run simulations for a limited amount of time and consider only limited scenarios. Unsplash
This is the hidden bottleneck in climate science. Because each simulation costs so much time, money, and computing power, researchers can only run a handful of scenarios. They cannot easily ask "what if" questions or explore many possible futures, because every run ties up a supercomputer for weeks. One of the researchers explained it simply: physics-based climate models solve equations about the atmosphere and oceans and run on supercomputers to make predictions for tens or even 100 years. But running such a physics-based model is very expensive in terms of time, compute, and energy. Unsplash
The AI approach works completely differently. Instead of solving the underlying physics equations step by step, which is what makes traditional models so slow, the AI learns the patterns of how the climate behaves and then generates predictions. The researchers call their model a "machine learning emulator." It emulates the climate rather than calculating every equation from scratch. The genius is that it does this without sacrificing much accuracy. In addition to running much faster than the state of the art, the model is also nearly as accurate without being anywhere near as computationally expensive. UnsplashUnsplash
ScienceDaily's coverage of the research: https://www.sciencedaily.com/releases/2024/12/241202150154.htm
What It Could Unlock
The exciting part is not just that this one model is fast. It is what becomes possible when climate modeling stops requiring a supercomputer. The researchers developed a generative AI climate prediction model that is fast and agile enough to be used as a tool not just by scientists, but by anyone whose decisions are affected by climate trends. Unsplash
Think about who that includes. Farmers deciding what to plant. Cities planning for floods and heat waves. Insurance companies pricing risk. Governments designing infrastructure. Energy companies planning for the grid of the future. All of these decisions depend on understanding how the climate will behave, and until now, the detailed modeling that informs those decisions has been locked away inside a small number of supercomputing centers. A model that runs on affordable hardware could put climate prediction in the hands of far more people who need it.
It also means scientists can finally explore many more scenarios. Instead of running one or two expensive simulations, they can run dozens, testing different assumptions and possible futures, which produces a much richer and more useful picture of what might actually happen. This is part of a broader wave of AI accelerating science. Researchers have shown climate models that run about 25 times faster by combining generative methods with physics-based data, as part of a broader trend where AI wins when it is tied to a real scientific job rather than used as a novelty. Wikipedia
UC San Diego's roundup of AI-enabled breakthroughs: https://today.ucsd.edu/story/nine-breakthroughs-made-possible-by-ai
What This Means For You
Even if you never run a climate model, this story matters for a few reasons.
First, it is a clear example of AI doing genuine good in the world, away from the hype. Most AI news is about chatbots, valuations, and corporate drama. This is AI being used to help humanity understand and prepare for one of the biggest challenges we face, and it is working. The same technology that makes funny images can also help a coastal city plan for rising seas. That is worth remembering when the AI conversation gets cynical.
Second, it points to a future where understanding the planet gets dramatically more accessible. Better, faster, cheaper climate prediction means better decisions about everything from where to build to how to farm to how to protect vulnerable communities. The benefits flow downstream to all of us, even if we never see the model itself.
Third, it is a preview of how AI is going to transform science across the board. The same trick used here, training an AI to emulate an expensive physics simulation, is being applied to drug discovery, materials science, weather forecasting, and dozens of other fields. The pattern is consistent: take a problem that used to require months of supercomputer time, and let an AI learn to approximate the answer in hours. As this approach spreads, the pace of scientific discovery itself speeds up. We are watching the early days of AI becoming one of the most powerful tools science has ever had, and the planet may be one of the biggest beneficiaries.
We will keep tracking the ways AI is accelerating real science and bring you the next breakthrough as it lands. Stay curious out there.
10x the context. Half the time.
Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.
10x the context. Half the time.
Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.