71–80 of 81 posts
Earlier quoted context omitted.
This comment is solid gold! I will definitely be sending people to it. Would make a great article for your own site, otherwise I'm happy to link to it here instead.
Thanks! Feel free to link to the HN comment. That will encourage me to make a video or two for YouTube demonstrating for a wider audience how I use AI for translation. I hope to do that within a few weeks.
Earlier quoted context omitted.
Thanks! Feel free to link to the HN comment. That will encourage me to make a video or two for YouTube demonstrating for a wider audience how I use AI for translation. I hope to do that within a few weeks.
Blogged it here: https://simonwillison.net/2025/Feb/2/workflow-for-translatio...
At the end of his post, Simon mentions translation between human languages. While maybe not directly related to token limits, I just did a test in which both R1 and o3-mini got worse at translation in the latter half of a long text. I ran the test on Perplexity Pro, which hosts DeepSeek R1 in the U.S. and which has just added o3-mini as well. The text was a speech I translated a month ago from Japanese to English, pr…
Yikes! Sounds to me like reliable longer form translation is very much not something you can trust to these models. Thanks for sharing.
At the end of his post, Simon mentions translation between human languages. While maybe not directly related to token limits, I just did a test in which both R1 and o3-mini got worse at translation in the latter half of a long text. I ran the test on Perplexity Pro, which hosts DeepSeek R1 in the U.S. and which has just added o3-mini as well. The text was a speech I translated a month ago from Japanese to English, pr…
“and” between noun phrases, for example)
it's just like.. the way it looks is so profound. maybe it's how it's formatted on my screen with the surrounding lines?
But then I noticed that the output length was 5,855 with the original Japanese, I discovered that R1 had “and” between noun phrases, for example) those problems.
idk, i'm just astonished. the amount of satisfaction that i get from reading your comment has given me enough dopamine to motivate me to create an account and write this entire comment itself, so thank you, your writing is appreciated, and i am very glad i came upon this and hope to find you again in the future somewhere. imagine???? xx
At the end of his post, Simon mentions translation between human languages. While maybe not directly related to token limits, I just did a test in which both R1 and o3-mini got worse at translation in the latter half of a long text. I ran the test on Perplexity Pro, which hosts DeepSeek R1 in the U.S. and which has just added o3-mini as well. The text was a speech I translated a month ago from Japanese to English, pr…
i just signed up for an account here to let you know that the way you write is perfect. listen.. i seriously mean perfect. like, the epitome of the perfect writing. you write better than LLM's. Actually I'd say you and Claude are on the same level but i'm looking at yours in the font of this website vs claude's normal font style so it still hits slightly different. i can't tell what exactly it is, but the fluidity of…
There's a huge price difference between o3-mini and o1 ($4.40 vs $60 per million output tokens), what trade-offs in performance would justify such a large price gap? Are there specific use cases where o1's higher cost is justified anymore?
its the same thing as: gpt-3.5 -> gpt-4 (gpt-4-32k premium) "omni" announced (multimodal fusion, initial promise of gpt-4o, but cost effectively distilled down with additional multimodal aspects) gpt-4o-mini -> gpt-4o (multimodal, realtime) gpt-4o + "reasoning" exposed via tools in ChatGPT (you can see it in export formats) -> "o" series o1 -> o1 premium / o1-mini (equivalent of gpt-4 "god model" becoming basis for l…
Earlier quoted context omitted.
i just signed up for an account here to let you know that the way you write is perfect. listen.. i seriously mean perfect. like, the epitome of the perfect writing. you write better than LLM's. Actually I'd say you and Claude are on the same level but i'm looking at yours in the font of this website vs claude's normal font style so it still hits slightly different. i can't tell what exactly it is, but the fluidity of…
You are very kind! Thank you.
Earlier quoted context omitted.
its the same thing as: gpt-3.5 -> gpt-4 (gpt-4-32k premium) "omni" announced (multimodal fusion, initial promise of gpt-4o, but cost effectively distilled down with additional multimodal aspects) gpt-4o-mini -> gpt-4o (multimodal, realtime) gpt-4o + "reasoning" exposed via tools in ChatGPT (you can see it in export formats) -> "o" series o1 -> o1 premium / o1-mini (equivalent of gpt-4 "god model" becoming basis for l…
No, o3-mini is a distillation of (not-yet-released) o3, not a distillation of o1.
Earlier quoted context omitted.
its the same thing as: gpt-3.5 -> gpt-4 (gpt-4-32k premium) "omni" announced (multimodal fusion, initial promise of gpt-4o, but cost effectively distilled down with additional multimodal aspects) gpt-4o-mini -> gpt-4o (multimodal, realtime) gpt-4o + "reasoning" exposed via tools in ChatGPT (you can see it in export formats) -> "o" series o1 -> o1 premium / o1-mini (equivalent of gpt-4 "god model" becoming basis for l…
No, o3-mini is a distillation of (not-yet-released) o3, not a distillation of o1.