# ChatJimmy 100 Hacker News commenters recommended ChatJimmy and 24 criticized it, November 2025 to November 2026. It was recommended in 9 of 13 months. It meets the Hacker Approved badge rule. **Verdict:** Recommended as a startling demo of near-instant LLM inference on Taalas's dedicated model chip, though it runs an old small Llama model that gives inaccurate, hallucinated answers. ## Why people recommend it - Near-instant, extremely fast responses (15 comments): "Compaction, busting the cache, and other issues like that will lose significance when you're running at 15k tokens per second like chatjimmy." ([source](https://news.ycombinator.com/item?id=49212327)) "It's eye opening: generated an AVX-512 optimized Mersenne Twister in C in 0.076s, 13,706 tok/s." ([source](https://news.ycombinator.com/item?id=48663358)) "I remember looking at this earlier this year and being that fast just feels fake." ([source](https://news.ycombinator.com/item?id=48691248)) - Shows promise of model-on-silicon (6 comments): "things like chatjimmy, I sometimes wonder if we might have built in chips with specialized models" ([source](https://news.ycombinator.com/item?id=48730716)) "it's running against Taalas' "hardcore" silicon model, ie a dedicated, ASIC-like chip." ([source](https://news.ycombinator.com/item?id=48027907)) "Millisecond inference even on much older models will change the whole game." ([source](https://news.ycombinator.com/item?id=48030985)) - Fast and cheap beats smart (3 comments): "There is an emergent space for super-fast-models esp finetuned or guardrailed to solve very specific latency sensitive tasks." ([source](https://news.ycombinator.com/item?id=49626374)) "this seems like a better choice to use a tool like this so you don't have your agent burn tons of tokens" ([source](https://news.ycombinator.com/item?id=48717655)) "a lot of LLM use cases don't need smart, they need fast and cheap" ([source](https://news.ycombinator.com/item?id=47091425)) - Speed boosts reasoning and agents (3 comments): "Compaction, busting the cache, and other issues like that will lose significance when you're running at 15k tokens per second like chatjimmy." ([source](https://news.ycombinator.com/item?id=49212327)) "it's incredible. Both reasoning and tool use generation scale with TPS" ([source](https://news.ycombinator.com/item?id=49205020)) "this seems like a better choice to use a tool like this so you don't have your agent burn tons of tokens" ([source](https://news.ycombinator.com/item?id=48717655)) ## What people warn about - Inaccurate answers and hallucinations (9 comments): "That's really impressive from a speed point of view but it hallucinates like it's on drugs." ([source](https://news.ycombinator.com/item?id=48220965)) "Is super fast but also super inaccurate, I would say not even gpt-3 levels." ([source](https://news.ycombinator.com/item?id=47087869)) "Its a cool demo, but its gpt-3.5 level stupid, or worse." ([source](https://news.ycombinator.com/item?id=49206243)) - Old, small Llama 3.1 8B (5 comments): "this leads to super-fast speeds but the model they're using is an old small Llama model so the quality is pretty bad" ([source](https://news.ycombinator.com/item?id=47779976)) "It is to be fair running a pretty trash model, a quantized Llama 3.1 8B from ~18 months ago" ([source](https://news.ycombinator.com/item?id=47141679)) "it's based on a 2+ year old tiny model. It's very much an alpha proof of concept" ([source](https://news.ycombinator.com/item?id=49293494)) - Outdated knowledge cutoff (2 comments): "It's great at knowledge and basic stuff but the output is complete junk for anything concerning new facts" ([source](https://news.ycombinator.com/item?id=49204097)) "training material is stopped at around July 2022, a little too outdated" ([source](https://news.ycombinator.com/item?id=49205005)) - Limited proof-of-concept features (4 comments): "the combination of small context window, no tool calls and just small model, made me understand, that it's not going to work" ([source](https://news.ycombinator.com/item?id=48366527)) "Chat jimmy launched in february, and nothing has come of it, not even GA let alone improvements with intelligence." ([source](https://news.ycombinator.com/item?id=50232509)) "it's based on a 2+ year old tiny model. It's very much an alpha proof of concept" ([source](https://news.ycombinator.com/item?id=49293494)) ## Alternatives hackers recommend - [Claude](https://hackerapproved.com/p/claude.md): 5175 recommend - [Google Gemini](https://hackerapproved.com/p/gemini.md): 2195 recommend - [ChatGPT](https://hackerapproved.com/p/chatgpt.md): 2186 recommend - [Grok](https://hackerapproved.com/p/grok.md): 727 recommend - [Microsoft Copilot](https://hackerapproved.com/p/microsoft-copilot.md): 139 recommend ## Where to get it - [Visit the official site](https://chatjimmy.ai/) ## Cite as Hacker Approved, "ChatJimmy", https://hackerapproved.com/p/chatjimmy-ai/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).