# BERT

19 Hacker News commenters recommended BERT and 4 criticized it, November 2025 to November 2026. It was recommended in 10 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Fast, cheap model that people fine-tune for classification and sentiment analysis, often competitive with larger LLMs. Caveat: it needs good data and expertise.

## Why people recommend it

- Good results when fine-tuned (6 comments): "they do a fairly decent job as long as you have good data and know what you're doing" ([source](https://news.ycombinator.com/item?id=45926500)) "you could just fine-tune BERT and get a faster model with better performance" ([source](https://news.ycombinator.com/item?id=48631740)) "Which in my experience is indeed quite effective (and easy/fast)." ([source](https://news.ycombinator.com/item?id=45926901))
- Small models beat large LLMs (3 comments): "its way more useful to train small purposeful models than to use an entire frontier LLM" ([source](https://news.ycombinator.com/item?id=50322016)) "A small transformer like BERT or variants is a better fit." ([source](https://news.ycombinator.com/item?id=48630182)) "We had good small language models for decades." ([source](https://news.ycombinator.com/item?id=47204374))
- Fast and cheap to run (3 comments): "Something like BERT could be and has been tuned to provide similar "decision making" at a similar latency and cost advantage" ([source](https://news.ycombinator.com/item?id=49753246)) "you could just fine-tune BERT and get a faster model with better performance" ([source](https://news.ycombinator.com/item?id=48631740)) "perform quite well in embedded applications" ([source](https://news.ycombinator.com/item?id=49868556))
- Strong at classification and sentiment (3 comments): "BERT was highly usable for classification and sentiment analysis a whopping 9 years ago" ([source](https://news.ycombinator.com/item?id=49768373)) "already known as SotA for text classification and similarity back in 2023" ([source](https://news.ycombinator.com/item?id=45926480)) "lightweight encoder models like BERT are great at sentiment analysis" ([source](https://news.ycombinator.com/item?id=45796260))

## What people warn about

- Too weak for real use (2 comments): "it was only good enough to be worth investigating further, not to actually use" ([source](https://news.ycombinator.com/item?id=47585162)) "You couldn't really get it to do useful stuff" ([source](https://news.ycombinator.com/item?id=49767752))

## Alternatives hackers recommend

- [Claude Opus](https://hackerapproved.com/p/claude-opus.md): 3331 recommend
- [OpenAI GPT](https://hackerapproved.com/p/gpt-5-5.md): 1205 recommend
- [Claude Sonnet](https://hackerapproved.com/p/claude-sonnet.md): 854 recommend
- [DeepSeek V4 Flash](https://hackerapproved.com/p/deepseek-v4-flash.md): 828 recommend
- [Fable](https://hackerapproved.com/p/fable.md): 823 recommend

## Cite as

Hacker Approved, "BERT", https://hackerapproved.com/p/bert/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).