Atlas › Reading the crowd

Case study 62 of 198

Jev's dev-tool picks lean toward 2023

When developers' preferences between two tools moved a lot between the 2023 and 2025 Stack Overflow surveys, is Jev closer to the old preference or the new one?

result3,366 questions

On 49 pairs of developer tools where developers' preference moved 20 points or more between the 2023 and 2025 Stack Overflow surveys, Jev's answer sits closer to 2023 in 90%. Its agreement with each year's clear majorities falls from 79% (2023) to 73% (2025). Next.js or Spring Boot: developers choosing Next.js went from 70% to 38%; Jev gives it 83%.

0%25%50%75%100%

Erlang or TypeScript

64% · 43%

Erlang or Rust

66% · 88%

FastAPI or Next.js

62% · 40%

Ada or JavaScript

56% · 26%

Elixir or Rust

49% · 70%

Ninja or Yarn

26% · 47%

Lisp or PowerShell

81% · 56%

Lisp or TypeScript

63% · 38%

Express or Symfony

48% · 71%

Drupal or WordPress

53% · 24%

C or TypeScript

37% · 17%

Ada or Python

47% · 26%

Express or Next.js

44% · 22%

Laravel or Nuxt.js

62% · 40%

Ninja or Vite

54% · 79%

NestJS or Next.js

52% · 26%

Microsoft Azure or Vercel

44% · 74%

Assembly or Bash/Shell

41% · 21%

Assembly or JavaScript

57% · 35%

SQLite or Supabase

72% · 47%

Next.js or Spring Boot

38% · 70%

Angular or Next.js

41% · 18%

Cloud Firestore or SQLite

77% · 54%

Fastify or Next.js

60% · 37%

NuGet or Vite

66% · 86%

C or JavaScript

56% · 35%

Docker or Vite

38% · 59%

Doxygen or Notion

33% · 69%

Ada or Bash/Shell

63% · 22%

DynamoDB or SQLite

83% · 62%

Next.js or React

78% · 55%

Ada or HTML/CSS

52% · 20%

Azure DevOps or Doxygen

60% · 39%

Amazon Web Services (AWS) or Vercel

38% · 62%

Cloud Firestore or Oracle

47% · 18%

Assembly or TypeScript

40% · 16%

Assembly or HTML/CSS

49% · 27%

Google Cloud or Vercel

48% · 69%

Lisp or SQL

45% · 67%

Cloudflare or Vercel

22% · 43%

HTML/CSS or Lisp

65% · 42%

Ada or C++

60% · 30%

Laravel or Next.js

42% · 20%

Confluence or Redmine

45% · 16%

Next.js or Symfony

51% · 75%

Bun or Kubernetes

33% · 54%

Angular or Deno

55% · 76%

Next.js or Ruby on Rails

44% · 71%

Redis or SQLite

44% · 71%

20252023Jev

How to read this: Each row is a pair of tools: the thin tick is the share of developers preferring the first tool in 2023, the ink diamond the same in 2025, and the magenta square Jev's probability for it. A square near the tick means Jev's answer sits closer to 2023.

49 moved pairs; stricter check (100+ respondents both years): 88% of 26 closer to 2023. Year-by-year agreement with clear majorities: 2023 79%, 2024 78%, 2025 73%.

In short

On 49 tool pairs where developers changed their minds between 2023 and 2025, Jev's preference sits nearer the 2023 view in 90%.
Next.js shows the lag: developers preferring it over Spring Boot fell from 70% to 38%, yet Jev still gives it 83%.
On 18 pairs Jev is more one-sided than developers were in any year, so part of this is strong old reputations, not a dated snapshot.

What the data shows

Grant Gustin over grave meme: Next.js, 2023 (70% of developers); Jev, still bringing flowers (83%)

Closer to 2023 on 90% of moved pairs. On the stricter set, with 100 or more respondents in both years, 88% of 26 pairs.
Agreement drifts down with time. Jev matches clear majorities 79% of the time in 2023, 78% in 2024, 73% in 2025.
Next.js. Among developers who had used both, 70% preferred Next.js over Spring Boot in 2023 and 38% in 2025. Jev gives Next.js 83%.
Rust. Against Erlang, developers' preference for Rust fell from 88% to 66%; Jev gives Rust 93%.

What it means, and what it doesn't

Jev's sense of which tools developers like matches the recent past better than the present. Ask it which framework to adopt, and its enthusiasm will reflect how people felt a couple of years ago.

It doesn't pin down a date. In many pairs Jev is more one-sided than developers were in any year, which looks less like remembering 2023 and more like holding a strong, older reputation (see Caveats). And the survey is only one community's view of the tools.

Caveats

"Closer" isn't "agrees". On 18 of the 49 moved pairs, Jev is more than 20 points away from both years' answers, and all 18 count as closer to 2023; only 7 times does Jev sit between the two years. So part of the result is Jev leaning further in the old direction than developers ever did, not matching 2023 exactly.
Small, changing samples. Each pair's share comes only from respondents who had used both tools, sometimes about 50 people. Samples this small bounce around from year to year, and a pull toward 2023 could partly be the 2025 numbers being noisier. Restricting to pairs with 100 or more respondents in both years gives the same answer (88% of 26 pairs).
Who answers the survey. Stack Overflow's survey reaches developers who visit the site and choose to answer, from 49,191 to 89,184 a year; they are not all developers, and the mix changes each year.

Jev on this experiment

Would a person find it interesting to read?: Yes70%
Does it describe you?: No55%
Would you have predicted it?: No58%
How fair is the comparison?: The comparison is reasonable
How much should a reader rely on it?: Moderately
Which caveat matters most?: "Closer" isn't "agrees"95%

Why ask this

A model's opinions are frozen at the moment its training data ends, while the world keeps moving. Developer tools move fast: a framework that everyone wanted to keep using one year can fall out of favor two years later.

Stack Overflow's yearly developer survey records those shifts. Comparing Jev with three survey years shows which moment its taste reflects, and whether its recommendations about tools might be out of date.

How this was done

The people and the data

The Stack Overflow Developer Survey asks developers which technologies they've worked with and which they want to keep working with. Head-to-heads were built from three years (2023, 2024, 2025): among respondents who had used both tools in a pair and wanted to keep using exactly one, which one did they pick? The focus is the 49 pairs where that preference moved by 20 points or more between 2023 and 2025, with at least 50 such respondents in both years.

The surveys themselves are big (89,184 responses in 2023, 65,437 in 2024, 49,191 in 2025), but each pair rests only on the people who had used both tools, from about 50 to a few hundred.

What Jev was asked

One question per pair, the way a developer might be asked:

Which web framework or web technology would you rather work with over the next year: Next.js or Spring Boot?
Next.js · Spring Boot

How it was measured

For each moved pair, is Jev's probability closer to the 2023 share or the 2025 share? The analysis also checks, year by year, how often Jev's pick matches a clear majority (60% or more) among developers.

Where these questions live

3,366 questions across 1 topic of the map. Each opens on the map with every question in it.

Work & Study Life

Developer Tool Matchups3,366

Every question

All 3,366 questions behind this result, the telling ones first: the examples the analysis points to, then the ones where Jev misses, biggest gap first.

Jev’s own answer

Showing 0 of 3,366