Writing AI was the opening act. Jasper, Notion AI, a dozen similar tools, these were the products that were going to unlock AI for every marketer, copywriter, and small business owner who couldn't code. The promise was democratic: AI for people who don't program.
Our current scoreboard tells a different story.
The Category Nobody Wants to Discuss
Jasper sits at 7.6. Last place. In a field of ten tools, the platform that helped pioneer the "AI writing assistant" category for non-technical users landed at the bottom. Notion AI does better at 8.4, but that score puts it tied with Runway (video generation) and GitHub Copilot. Those tools are solving considerably harder technical problems.
The gap between the top and bottom of our board is 1.4 points: Cursor at 9.0, Jasper at 7.6. In a scoring band where most tools cluster between 8.1 and 8.6, that spread is not small. It's the widest distance on the entire list.
Three Code Editors, One Dominant Winner
The top of our scoreboard looks like a developer's desktop.
Cursor leads at 9.0, a full 0.4 points ahead of anything else ranked. Windsurf follows at 8.5. GitHub Copilot holds at 8.4. Three code-editing tools in a list of ten, occupying positions one, four, and five. The "AI coding revolution" was supposedly going to arrive last, after the easier wins. According to our numbers, it arrived first and pulled the furthest ahead.
Why Writing Underperforms
The writing category's position at the bottom is not really about bad products. It's about what AI is structurally good at.
Code has hard constraints. A function compiles or it doesn't. An autocomplete suggestion is syntactically valid or it gets rejected immediately. AI excels at problems with clear, verifiable answers, and code is dense with them. Cursor's 9.0 reflects this: a constrained problem solved extremely well.
Writing has no such constraints. "Is this paragraph good?" has no binary answer. AI writing tools must clear a subjective bar, and the gap between "grammatically correct" and "actually useful to a working writer" is exactly where Jasper loses points. This is not a critique of Jasper specifically. It is a structural observation about what the category asks AI to do.
Notion AI at 8.4 outperforms Jasper meaningfully, but Notion's advantage comes partly from being embedded inside an existing productivity workflow. People use Notion AI because they already live in Notion, not purely because the writing output is superior. That's a context advantage, not a writing quality win.
The Voice Category Has One Representative
One detail worth flagging: ElevenLabs scores 8.6, tied with Perplexity AI for second overall, and it is the only voice or audio tool on our board. Voice synthesis is arguably more technically demanding than writing assistance, and arguably more transformative commercially, yet the category shows up with a single entry.
ElevenLabs earned that 8.6 without competition from within its own space. Whether that reflects genuine category-level quality or a shortage of mature alternatives is something to revisit when the next round of audio tools matures enough to rank.
For now, the pattern is clear. Tools built for hard, specific, measurable problems lead the scoreboard. Tools built for subjective, open-ended creative work trail behind. That structural gap is unlikely to close on its own.