I just submitted to ICLR. As of last night’s abstract deadline, there were over 60,000 submissions registered. For comparison, last year ICLR started out with 19,525 submissions as per the stats posted on their blog. 779 (3.99%) were desk-rejected, 5,042 (25.82%) were withdrawn, and 13,763 papers — 70.49% of the total number of submissions — ultimately reached a final decision. Assuming the trends remain the same, we can expect this year’s cycle to yield approximately 42,300 valid submissions that go through peer review, while the number of reviewers in the eligible pool remains roughly the same. The practice of doing and publishing research has evidently fundamentally changed. When I started my research career more than a decade ago, we would spend weeks, sometimes months, trying to understand a problem space before forming a mature point of view on what was worth working on. That process involved reading backwards through the literature, figuring out why successive generations of methods made the design choices they did — often by re-implementing them on your problem and specific dataset, with much struggle to reproduce hyperparameter choices so the comparison was apples-to-apples — to eventually developing something that sticks, and generalizes nicely to multiple contexts, domains, and across architectures. Now, with LLM tooling, the research loop has compressed dramatically. We delegate these tasks to autonomous research agents who do literature surveys for us, generate and implement experiment ideas for us, and prepare the final writeup for us — leaving us to ramp up to new domains in record time, but never fully develop a deep understanding of the problem landscape. We aren't assessing novelty truthfully, because we don't write or read papers the way we used to, and rarely pay attention to anything published more than a year ago. Developing context on how a problem space has evolved — what we like to call “research taste and wisdom” — is now a rare skill. Much of what people claim novelty for is the same idea packaged in different marketing language. Reminds me of one of those 90s lottery machines: a giant transparent sphere filled with numbered balls, each representing a research idea, all bouncing around at high speeds until one gets sucked into the draw tube, and emerges as the winning number. We take existing ideas, throw them into a massive lottery sphere, crank up the speed and ask, “what’s the next number?”, and voila! A ball pops up at the slot that looks sufficiently different to be given a new name, and a “novel” paper emerges. We obviously have an incentive problem as a field, and it extends far beyond researchers to the whole AI ecosystem. When claims of progress help determine where capital, compute, talent, and influence accumulate, research priorities become increasingly entangled with commercial interests. The longer, less diplomatic version of my views is on my blog: mariya.fyi/posts/research-lo… – at San Francisco, CA