Alright, let's just get this out there. I've been using Elicit for months to sift through research papers on infrastructure patterns and monitoring best practices. The old UI had a learning curve, sure, but once you got it, you could fly. You could set up complex queries, compare multiple result sets in different tabs, and export data in a way that made sense for further analysis.
The new "streamlined" UI feels like it was designed for someone who reads three papers a year. My main gripes:
* **Loss of multi-tab/workspace view:** I used to have my primary search in one tab, then open key papers in adjacent tabs to cross-reference methodologies. Now it's a single linear flow. It's like going from a multi-monitor Kubernetes dashboard back to a single `kubectl get pods` output stream.
* **Export functionality is buried and less granular:** Before, I could easily export a CSV of, say, 50 papers with their key claims, URLs, and publication years. Now the export options seem to default to a prettified, truncated view. The data feels locked in. It reminds me of when a SaaS monitoring tool replaces a detailed metrics table with a pretty, non-interactive chart.
* **Keyboard shortcuts are either gone or changed without clear indication:** Muscle memory is broken. This is a productivity killer for any power user. It's the equivalent of a major IDE release changing all the debug key bindings.
* **The dense information view is gone:** They've increased white space and font size, which means I'm scrolling more to see the same amount of information. I don't need a reading pane that feels like a mobile blog; I need a data-dense table that I can scan quickly.
Here's a concrete example of the workflow break. I was researching "eBPF for Kubernetes network monitoring." My old workflow:
1. Complex query in main tab.
2. Open 10 relevant papers in adjacent tabs.
3. Use the compare function to extract a table of tools vs. capabilities.
4. Export that table to CSV for documentation.
Now, I'm stuck with a sequential list. I have to open a paper, read it, go back, find my place, open the next one. There's no easy "at-a-glance" comparison matrix anymore. It's slowed my research process down by at least 30%.
It feels like a classic case of optimizing for new user acquisition at the expense of the core users who actually derive deep, repetitive value from the tool. Has anyone else run into these specific issues, or found workarounds? Am I just missing a "pro mode" toggle somewhere?
Automate everything. Twice.