# Trying out the new AI models
**URL:** https://meta.discourse.org/t/trying-out-the-new-ai-models/343123
**Category:** Community Building
**Tags:** ai
**Created:** [December 18, 2024, 7:21am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123 "2024-12-18T07:21:47Z")
**Posts on this page:** 10
**Page:** 1
### Author:  [@Don](https://meta.discourse.org/u/Don)
#### Post date: [December 18, 2024, 7:21am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/1 "2024-12-18T07:21:47Z")
Hello đź‘‹
I make this topic for share experiences with the new AI models using on Discourse.
I set up a few days ago the `grok-2-1212` for topic summaries. It works really well. The language detection and quality is well enough. I tried it also with AI bot but it most of time failed I guess it can’t handle the tools well yet.
I also tried the Gemini Flash 2 for AI bot. It works fast and add great answers however sometimes it seems it can’t handle well the tools and broke the answer. Sometimes it is a simple markdowns formatting issue, sometimes it can’t search. On my forum most of time says didn’t find anything on the forum but I know more topic about that subject…
---
### Author:  [@MihirR](https://meta.discourse.org/u/MihirR)
#### Post date: [December 18, 2024, 7:48am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/2 "2024-12-18T07:48:12Z")
> [@Don](#):
>
> On my forum most of time says didn’t find anything on the forum but I know more topic about that subject…
For the search problem, do you think it could be related to the AI not fully indexing the forum content or there might be a mismatch in query understanding?
---
### Author:  [@Don](https://meta.discourse.org/u/Don)
#### Post date: [December 18, 2024, 8:00am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/3 "2024-12-18T08:00:31Z")
If have no idea. Most of time it searches for nothing “” or failed with timeout… but sometimes do the search correctly and linked the correct topics. It will be good I think but it’s strongly experimental yet.
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 18, 2024, 8:15am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/4 "2024-12-18T08:15:43Z")
Have you tried xml tools? I found that on grok they work quite well
---
### Author:  [@Don](https://meta.discourse.org/u/Don)
#### Post date: [December 18, 2024, 8:28am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/5 "2024-12-18T08:28:14Z")
Thanks, I tried it now. Yeah looks better, the problem is mostly happening now when I create a new conversation. It starts in English something like: _I am searching for…in sitename…_ and stop replying. Sometimes it continue the answer after the English _I am searching…_ sentence on the correct Hungarian language and add good answer. However if I reply the grok response after that in the conversation it will works great.
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 18, 2024, 8:32am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/6 "2024-12-18T08:32:54Z")
This is really interesting, I kind of want to allow “grounding” examples as an option for personas, it could totally solve this
---
### Author:  [@MihirR](https://meta.discourse.org/u/MihirR)
#### Post date: [December 18, 2024, 8:36am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/7 "2024-12-18T08:36:38Z")
Bingo! But do you think it’ll slow things down, especially with a lot of data? Could it affect response times for AI queries or search results, or is it all good?
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 18, 2024, 8:42am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/8 "2024-12-18T08:42:55Z")
The big problem examples have are “contamination”
The model learns shape, but also can mistakenly think a user said something they did not.
Ideally carefully crafting system messages can do the trick , that would be my first resort
Examples in a system message can lead to less leakage cause it can be clearer to a model it is just an example
A minimal thing I would recommend Don, is writing your system message in Hungarian, it could help
Maybe even try giving an xml tool example or two in the system message?
---
### Author:  [@MihirR](https://meta.discourse.org/u/MihirR)
#### Post date: [December 18, 2024, 8:48am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/9 "2024-12-18T08:48:02Z")
This genuinely sounds good, thanks for sharing : )
---
### Author:  [@Don](https://meta.discourse.org/u/Don)
#### Post date: [December 18, 2024, 9:08am UTC](https://meta.discourse.org/t/trying-out-the-new-ai-models/343123/10 "2024-12-18T09:08:55Z")
> [@sam](#):
>
> A minimal thing I would recommend Don, is writing your system message in Hungarian, it could help
I tried it but same result with `grok-2-1212` then I switched it to `grok-beta` and it works perfect but it works with English system message too…