The front page of AI.
AI has become impossible to keep up with.
Every day, AI companies release hundreds of new products, features, and models. Research labs publish breakthroughs, developers ship new tools, and open-source projects introduce capabilities that didn’t exist the day before.
In 2025 alone, 280,000 research papers and preprints were submitted to arXiv. Add the millions of repositories, model updates, benchmarks, and announcements, and the volume of information is beyond what any person or team can follow.
And it’s scattered everywhere.
Research lives on arXiv, code on GitHub, and models on Hugging Face, while product announcements are spread across company blogs, changelogs, and social media. Each platform shows only a fraction of what’s happening.
Keeping up means checking dozens of websites, following hundreds of accounts, and relying on algorithms that prioritize engagement over relevance. Much of the work published in AI never reaches a social feed at all.
So researchers spend months solving problems that already have answers, and engineers rebuild tools someone else has already shipped. Companies make technical decisions without knowing that a new release has changed their options.
| Year | Submissions a month |
|---|---|
| 2010 | 5,830 |
| 2011 | 6,380 |
| 2012 | 7,050 |
| 2013 | 7,720 |
| 2014 | 8,120 |
| 2015 | 8,750 |
| 2016 | 9,420 |
| 2017 | 10,330 |
| 2018 | 11,750 |
| 2019 | 12,920 |
| 2020 | 14,830 |
| 2021 | 15,080 |
| 2022 | 15,500 |
| 2023 | 17,330 |
| 2024 | 20,080 |
| 2025 | 23,330 |
| 2026 | 30,045 |
One place to keep up with your field.
Every major industry has a place where its people go to understand what’s happening. Finance has its terminals and wire services. Politics and business have their front pages. AI is now reshaping all of them, yet the people building it still piece the field together from scattered tabs and feeds.
And nobody in AI follows all of it. A researcher in robotics, an engineer shipping voice agents, and a CTO choosing a model provider each care about a different part of the field. What they need isn’t more news. They need the part that matters to their work, complete, in one place, the moment it happens.
That’s why we built AlphaSignal.
We cover the entire AI industry, from the largest labs and the top 150 AI companies to the research and open-source work that will shape what they build next. Each person gets a single place to keep up with their own corner of it, without searching across dozens of sites or hoping the right post finds them.
Years of learning what matters.
AlphaSignal started as a newsletter with a simple promise: show people in AI the developments worth their time, and leave out the rest. It became the most-read newsletter in the industry, with readers at the companies and labs building AI.
Putting each issue together meant reading the whole field first and deciding what deserved attention. Doing that by hand stopped being possible early on, so we built software to do it with us.
We spent years refining it. Along the way, it learned what our readers had been teaching us: the most important work is rarely the loudest. A quiet paper can outweigh a big launch, and a small repository can change how a whole team builds.
A newsletter could share only a sliver of what that engine found, and every reader got the same issue. Meanwhile, releases began arriving at all hours, and our audience’s work kept splitting into narrower specialties.
So we built a newsroom on the same foundation, one that never stops publishing and lets each person keep up with their own part of AI.
Our mission
Accelerate AI progress by helping the industry keep up with itself.
Progress in AI depends on more than new ideas. It depends on how quickly those ideas reach the people who can build on them. Today, breakthroughs go unnoticed, work gets duplicated, and knowledge stays trapped on disconnected platforms.
We want to change that by becoming the center of AI: the place where researchers, engineers, developers, and technical leaders come to see what the field has accomplished and what’s coming next.
When the people building AI can see each other’s work as soon as it appears, every discovery reaches the next one faster. That’s the industry we’re building for.
AlphaSignal