AI talk: how far are you guys into it?

[Elmage G Ace]

You sound really convincing ,almost like ai trying to convince everyone that they are not going to take their jobs.

(Dale gribble…..rusty shakleford)

Lol!
With the rise of agents, it's inevitable that change is coming, given its potential, and just like with other cases where new technologies of such magnitude have been introduced, if you don't embrace or at least adopt it, you'll get left behind

 

[Tennis Robot]

I'm tired of LLMs. Not that this space is necessarily of interest to me, but I think that the recent academic fixation on LLMs is misguided, given their flaws. Last year I spent some time working on a research project aiming to use discrete diffusion models to achieve similar results. However, I ultimately disagreed with this method as well. While there is potential upside to using diffusion models - they don't need to predict tokens in order - they have other shortcomings (for instance, needing to determine the length of the output in advance).

Whatever the future of generative models is though, they certainly aren't going away. In my day-to-day life, I find them very useful for checking my understanding of a problem or a concept: I will often post a screenshot of a proof/solution I am reading through or trying to digest and ask if my interpretation of what is happening or my understanding of the significance of the result is correct. It is helpful/accurate about 80-90% of the time.

 

[The Big Boss]

I'm tired of LLMs. Not that this space is necessarily of interest to me, but I think that the recent academic fixation on LLMs is misguided, given their flaws. Last year I spent some time working on a research project aiming to use discrete diffusion models to achieve similar results. However, I ultimately disagreed with this method as well. While there is potential upside to using diffusion models - they don't need to predict tokens in order - they have other shortcomings (for instance, needing to determine the length of the output in advance).

Whatever the future of generative models is though, they certainly aren't going away. In my day-to-day life, I find them very useful for checking my understanding of a problem or a concept: I will often post a screenshot of a proof/solution I am reading through or trying to digest and ask if my interpretation of what is happening or my understanding of the significance of the result is correct. It is helpful/accurate about 80-90% of the time.

Are you tennis ai ?

 

[Joon]

I use it for BS reasons and I don't feel like using Google. Nothing important.

 

[Caliburn]

I partially outsourced my Google queries to it. Why Google it using Boolean Operators with crappy results if AI can provide better output while giving very specific input?

A while ago though, the IT team of where I work was developing an internal AI app for which I was one of the test persons. Currently, our IT environment is still very strongly SharePoint based and I thought it could be useful in finding and organizing documents that are located all over the place, but the results were poor.

I haven't heard anything about in a while now, so I'm pretty sure they're too busy for their little side project and it's somewhere in the freezer.

 

[Kagutsuchi]

We use it for Design / Product Videos / Social Media / Content .. Q/A and alot of stuff.

 

[Vroqrak]

Lately AI is everywhere, from people messing around with image generators to entire companies rebuilding their workflows around it. I was wondering how it looks for everyone here on AB. Are you already using AI tools in your daily life or work? If so, what’s the actual value you’ve gotten out of it?

For context: I work as an AI Architect for large tech companies, helping them how to actually apply AI instead of just hyping it up. That means a lot of building with agents, training models on company-specific data, and designing setups that don’t just sound cool but really deliver.

Curious to hear what you all are doing. Are you experimenting with prompts, coding with AI, maybe automating parts of your work, or still just watching from the sidelines?

ChatGPT is the largest funder of the leading genocidal nation in the world, I try not to use it.