"Losing relevance" is multi-dimensional. For a language to be rapidly losing relevance, you typically see:

• Usage decline in ecosystem metrics (GitHub activity, npm downloads, TIOBE/PYPL share).

• Ranking drops in composite indices (RedMonk, TIOBE).

• Shrinking community and hiring demand (developer surveys, job postings).

• Platform shifts (e.g., Apple moving from Objective‑C to Swift).
  Based on the collected evidence through early 2026, a small set of languages stand out as clearly losing relevance faster than the rest, especially for new greenfield development, even though many will persist for years in legacy codebases.

Languages Losing Relevance Fastest (2023–2026)

1. CoffeeScript – From "ES6 before ES6" to Near-Obsolete

Why it's high on the list

Multiple analyses of "obsolete by 2026" languages note that CoffeeScript is "all but obsolete" and "on life support."
One such analysis reports that npm downloads for CoffeeScript had collapsed by about 92% from peak levels by 2023[1]. That is an extreme, quantifiable falloff in real-world use.
Key drivers

ES6+ JavaScript caught up and surpassed it:
Features that made CoffeeScript attractive—arrow functions, destructuring, comprehensions, cleaner syntax—arrived in mainstream JavaScript via ES6 and later.
Modern toolchains (Babel, TypeScript, modern bundlers) eliminated the need for a separate CoffeeScript-to-JS transpile step.
Ecosystem moved on:
Major frameworks (React, Vue, modern Node backends) standardized around modern JS/TypeScript, not CoffeeScript.
Even the CoffeeScript creator has encouraged developers to move to modern JavaScript in practice.
Implications

For new projects, CoffeeScript should be treated as effectively deprecated.
For legacy code, plan a gradual migration to TypeScript or modern JavaScript, ideally aligning with a broader refactor.

1. Objective‑C – Swift's Displaced Predecessor

Evidence of decline

The RedMonk January 2026 rankings show Objective‑C at #18, down from a long‑standing position around #9–10 since 2012[2].
RedMonk notes that, given this trajectory and competing languages around it, it's plausible Objective‑C may make a permanent exit from the Top 20 soon[2].
JetBrains' ecosystem surveys (2023–2025) report Objective‑C used by fewer than ~2% of developers, emphasizing its niche legacy status for Apple platforms.
Key drivers

Swift's dominance:
Apple's Swift is now the standard for new iOS/macOS development, with App Store statistics showing the vast majority of new apps using Swift.
Tutorials, official documentation, and sample code from Apple emphasize Swift.
Tooling & ecosystem shift:
New language features, concurrency models, and frameworks are optimized for Swift.
Hiring pipelines and bootcamps teach Swift; Objective‑C is no longer the default entry point.
Implications

Objective‑C remains relevant for maintaining large, legacy iOS/macOS codebases, but:
For greenfield work, using Objective‑C is hard to justify.
Teams should expect shrinking talent pools and fewer third‑party libraries over time.
Adoption risk is high; plan progressive interoperability and migration to Swift where feasible.

1. Ruby – From Rails Hype to Narrowed, Mature Niche

Evidence of decline

Tiobe/InfoWorld analysis (April 2025) highlights Ruby, Kotlin, and Swift dropping out of Tiobe's Top 20, with Ruby at 24th place after years as a stable top‑20 language[3].
Tiobe's CEO notes that these languages "held very stable top 20 positions until recently… but now seem to have lost traction and are likely to go out of fashion."[3]
Other commentary (JetBrains, ecosystem blogs) frames Ruby as declining in job postings and new‑project adoption relative to Python, Node.js, and Go.
Key drivers

Rails is no longer the default web choice:
New backends increasingly favor Node/TypeScript, Go, or JVM stacks for perceived scalability, ecosystem, and hiring pool.
Python's broad dominance:
Python has taken over the "productive scripting & backend + data" slot; Ruby is rarely chosen for AI/ML, and less so for greenfield web in larger organizations.
Platform perception:
Articles and commentary describe Ruby as "sinking in popularity, buried by Python" and surviving largely because of existing systems (Shopify, GitHub's legacy stack, internal tools) rather than new demand.
Implications

Ruby will persist for many years—Rails apps won't vanish—but:
For career planning, Ruby alone is increasingly risky unless you anchor in a strong employer or niche.
For new products, you sacrifice ecosystem momentum and hiring flexibility compared to TypeScript, Go, or Python.
Ruby is not "dead", but it is clearly losing share and mindshare relative to its 2010s peak.

1. Legacy Enterprise & Statistical Scripting Languages Under Pressure

This category shares a common pattern: they keep significant installed base but show clear long‑term erosion in popularity metrics.

4.1 Perl

Signals

Historical leader in scripting, but:
Multiple recent commentaries describe Perl as an "aging powerhouse losing its relevance" and a "declining powerhouse"[1][4].
As far back as classic Stack Overflow analyses, Perl was both one of the most‑disliked languages and among the fastest‑shrinking tags by question traffic (though that's older data).
TIOBE commentary in 2026 notes Python and R rising, with "long‑established alternatives continue to lose momentum," in context that has historically included Perl.
Drivers

Fragmented evolution and ecosystem competition:
Python, Ruby, and later Node/TypeScript displaced Perl for web and scripting.
Internal fragmentation (Perl 5 vs 6, now Raku) and slower modernization hurt mindshare.
Perception:
Many engineers now see Perl as legacy and hard to read/maintain compared to modern counterparts.
Implications

Perl persists in text processing, legacy systems, and some sysadmin tooling.
New adoption is minimal; for most organizations, migrating off Perl as you touch legacy systems is the lower‑risk strategy.

4.2 MATLAB and "Classic" Statistical Languages (outside R)

Commentary across surveys and indexes suggests MATLAB is flirting with dropping out of some top‑20 rankings as data science consolidates around Python and, to a lesser extent, R.
Where earlier analyses observed MATLAB among shrinking tags despite not being strongly "disliked," the trend today continues:
Universities and industry train more heavily on Python/R.
Open‑source ecosystems (NumPy, SciPy, PyTorch, TensorFlow) erode the need for a proprietary MATLAB stack in many domains.
Implications

MATLAB will remain for specialized engineering and embedded workflows, but new data teams overwhelmingly choose Python/R.
For institutional strategy, investing in MATLAB for general‑purpose analytics is increasingly hard to justify.

1. "Declining Hype" but Still Viable: Scala, Classic PHP, Older Java Styles

These languages are not collapsing the way CoffeeScript or Objective‑C are, but show relative decline for greenfield development decisions.

Scala
A 2026 analysis categorizes Scala explicitly as "📉 Declining hype", noting:
Important historical role (Spark, JVM data tooling, Twitter, LinkedIn, Airbnb).
Heavy toolchain, long build times, and sharp learning curve.
Many teams "that wanted pragmatic backend work now choose Java, Kotlin, Go, or TypeScript instead"[5].
Implication: Scala is steadily ceding new‑project territory to Kotlin/Java and Go, especially outside Spark ecosystems.
PHP (as a primary backend choice)
Some surveys and commentaries show PHP usage still very high in web (WordPress, legacy apps), but:
PYPL/TIOBE long‑term views and industry blogs often call PHP the "biggest loser" in share over the last decade, dropping from >20% to low single digits in some indices.
The perception is that modern greenfield platforms more often choose Laravel only in specific niches; enterprise and startups skew to Node/TS, Go, Java, or .NET.
Java (interest, not usage)
An ITPro analysis cites a 13% drop in interest for Java versus a smaller 5.3% for Python, based on search and skills‑demand patterns[6].
Enterprise use remains enormous, but:
Java has lost mindshare for "exciting new projects" to languages like Kotlin, Rust, Go, TypeScript.
For junior developers and bootcamps, Java is less often the default first language.
Implications

These languages are not "dead"; their installed base and corporate investments are large.
The decline is relative:
They are losing share against more modern or specialized stacks.
For new developers, relying only on these for career durability is riskier than combining them with newer stacks.

Important Caveat: Stack Overflow Question Decline vs Language Decline

One 2026 analysis of Stack Overflow data found Rust and Go exhibiting some of the largest declines in Stack Overflow question volume after the rise of AI coding tools, alongside broad drops across languages.

This does not imply Rust and Go adoption are falling:
JetBrains and ecosystem reports show Rust and Go continuing to grow in adoption and remain among the most admired or in‑demand languages.
The Stack Overflow drop primarily reflects developers getting answers from AI tools instead of Q&A sites, not language abandonment.
Takeaway: Do not interpret reduced Stack Overflow traffic alone as a sign a language is losing relevance; cross‑check with GitHub, hiring data, and survey adoption.

Practical Takeaways

For engineers and technical leaders:

Strong "avoid for new projects" list (unless you have a very specific legacy or niche reason):

CoffeeScript
Objective‑C (for new Apple‑platform apps)
Classic Perl for new general‑purpose scripting/web
MATLAB for general data science (vs Python/R)
Use with caution for greenfield backends / data platforms:

Ruby (Rails still fine in the right context, but harder to hire & scale teams)
Scala (outside of existing Spark ecosystems; prefer Kotlin/Java/Go for many JVM/infra cases)
Older PHP‑heavy architectures where modern alternatives fit better
Migration priorities:

If you own significant Objective‑C or Ruby backends, start planning phased interop + migration:
Objective‑C → Swift
Ruby/PHP/Perl → TypeScript, Go, Java/Kotlin, or Python depending on domain
Do this opportunistically alongside feature work and security upgrades, not as big-bang rewrites.