Talkie
Shares rank 36 of 72 language models by recommenders
People recommend this chatbot-style model, trained only on pre-1931 text, as a distinctive window into the writing and views of its era. Critics warn its training data is contaminated with later text, so results may be unreliable.
People recommend this chatbot-style model, trained only on pre-1931 text, as a distinctive window into the writing and views of its era. Critics warn its training data is contaminated with later text, so results may be unreliable.
32 people gave an opinion
- recommend it
- 19
- criticize it
- 13
- approval share
- 59%
Recommended in 4 of 13 months
- November 2025: not recommended
- December 2025: not recommended
- January 2026: not recommended
- February 2026: recommended
- March 2026: not recommended
- April 2026: recommended
- May 2026: not recommended
- June 2026: not recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: not recommended
- October 2026: not recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 19 people recommend itNeeds 10 or more, met.
- 1.4 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 4 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Fun to try and explore
“I love this, such a cute little guy.”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “best used for cool hacks like this”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “The live-feed is super cool.”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” Authentic view of 1930s perspectives
“The pre-1931 cutoff date for this model is probably as close to the end of the optimistic age as it's reasonable to get.”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “This is so awesome they built this, I've wondered what an LLM only trained on pre-1950s data would look like!”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “It's much more useful for getting vibes / gist of perspectives of the day.”
Comment from 29 Apr 2026 in “Talkie: a 13B vintage language model from 1930” Useful control baseline for research
“It is cool. I find the idea of trying to understand whether these types of models can come up with things like General relativity”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “a great control baseline for studying what's actually learned from pretraining vs. scale”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” Fully open, copyright-free model released
“one of the only models fully unencumbered from copyright”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “actually releasing a model counts for a whole lot”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930”
What people warn about
Each mark is one comment
Training data contaminated with later text
“The fact that the model is contaminated doesn't mean that it won't behave mostly like a model from that era. But it casts doubt on everything done with this model”
Comment from 30 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “But as they acknowledge later, there are multiple major data leakage issues in their training pipeline”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “This model is contaminated in subtle ways that make me skeptical of the results.”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” - Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930”
- Comment from 28 Apr 2026 in “CS Professor: To My Students”
Hallucinates and drifts into nonsense
“it's trying to convince me that he was born in 1839, even after I told it he was born in 1889”
Comment from 29 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “it hallucinated a film from 1923 that takes place in the Antarctic”
Comment from 29 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “Then it riffs on that, drifting off into plausible nonsense.”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” - Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930”
- Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930”
Repeats biased or offensive views
“ask it about its opinion on "coloured people" and you will lose every ounce of skepticism”
Comment from 30 Apr 2026 in “Talkie: a 13B vintage language model from 1930” “So finally it is settled. The war was about state rights, not about slavery.”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930” Hard to host; no mobile support
“hosting talkie myself was a slog and I had problems using platforms”
Comment from 29 Apr 2026 in “A Podcast with Talkie, a 13B model trained only on pre-1931 text” “The chat does not work on mobile /tablet”
Comment from 28 Apr 2026 in “Talkie: a 13B vintage language model from 1930”
Other language models people recommend
The four most recommended other language models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Claude Opus | 3,331 | 2,746 | 55% | 13 | 4 mo156 owners | — |
| OpenAI GPT | 1,205 | 987 | 55% | 13 | 6 mo47 owners | — |
| Claude Sonnet | 854 | 586 | 59% | 13 | 6 mo46 owners | — |
| DeepSeek V4 Flash | 828 | 223 | 79% | 8 | 1 mo24 owners | Approved |
| TalkieThis page | 19 | 13 | 59% | 4 | — | — |
What people said
32 of the 34 comments that mention it, newest first. Each links to the original comment.
“Still seems like a pretty reasonable "proof of concept"”
“Talkie's writing is so distinctive that it is often classified as human by Pangram.”
“The fact that the model is contaminated doesn't mean that it won't behave mostly like a model from that era. But it casts doubt on everything done with this model”
“ask it about its opinion on "coloured people" and you will lose every ounce of skepticism”
“It's much more useful for getting vibes / gist of perspectives of the day.”
“it's trying to convince me that he was born in 1839, even after I told it he was born in 1889”
“Talkie's myopic viewpoint does not represent the most farsighted viewpoint, merely the average.”
“it hallucinated a film from 1923 that takes place in the Antarctic”
“hosting talkie myself was a slog and I had problems using platforms”
“The writing is way better than both ChatGPT and Claude.”
“The live-feed is super cool.”
“This is so awesome they built this, I've wondered what an LLM only trained on pre-1950s data would look like!”
“it said the aether was not disproved by the experiment”
“accidentally slipped through into the training data”
“The chat does not work on mobile /tablet”
“The pre-1931 cutoff date for this model is probably as close to the end of the optimistic age as it's reasonable to get.”
“a great control baseline for studying what's actually learned from pretraining vs. scale”
“best used for cool hacks like this”
“So finally it is settled. The war was about state rights, not about slavery.”
“one of the only models fully unencumbered from copyright”
“I love this, such a cute little guy.”
“how inaccurate the results are, what use is there to it?”
“This model is contaminated in subtle ways that make me skeptical of the results.”
“Then it riffs on that, drifting off into plausible nonsense.”
“It doesn't know about the Great Depression (1929-WWII).”
“actually releasing a model counts for a whole lot”
“But as they acknowledge later, there are multiple major data leakage issues in their training pipeline”
“Talkie isn't entirely pure either though”
“It is cool. I find the idea of trying to understand whether these types of models can come up with things like General relativity”
“everything is a file and local first, BYOK”
“native app uses Parakeet (v2 or V3) on iOS”
“incredibly fast and made for devs”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.