Large language model

It's very hard to know really.

We're at a point now where most people are familiar with LLMs.

They've tried them for various things.

There's been a lot of hype, but actually they're, they're not actually that good at quite a lot of things.

I think it's super interesting.

We've seen this transition over the last 18 months really as we've been going through the, this journey of the transition of the buyers on the LLMs, and just also seeing the trust of the answer in the LLM taking more and more into account.

And so we've talked a lot about the need to build your brand, because if you're not on that shortlist, you s- that's the way you still are able to get on that, and I advocate a lot for building your brand on G2, because if they're gonna end up going to that review site after the LLM, that's another chance to be seen and remind them to make sure that you're on the shortlist as well.

So it's a really critical part of brand building.

This is very different now.

Now you have to speak to both the human and the agent, 'cause they're both crawling, and they're both understanding your site and pulling data in real time.

So you have to earn those trust signals in those sources and frequency of the LLMs.

The frequency is key.

This is episode 311.

What are the latest trends uncovered in the 2026 Django Developers Survey? How are Django users employing LLMs in their development process? Christopher Trudeau is back on the show this week with another batch of PyCoders weekly articles and projects.

We discuss a summary by Will Vincent about this year's Django Developers Survey.

The survey draws on responses from nearly 3,500 developers across more than 40 countries.

Nobody is winning.

About a third each for Claude and ChatGPT, with Copilot falling behind at 23%.

60% of respondents are using a web browser to interact with LLMs, while 13% use local models.

When learning, Django's official site is still the number one resource with 67% of people using it, but AI has come on strong with just over half using AI to help learn Django, which has surged ahead of other things like YouTube and other resources that used to be the second place.

I don't think most people could explain to me how electric cars work, but plenty of people own Teslas.

Can someone really explain to me how LLMs work, but they buy all these AI bottleneck stocks? So I think the answer is, if they don't have any crypto And we've been advocating a 2% position for more than a decade.

Our original recommendation for a 2% position, okay, for the average account for Fundstrat that actually took our advice, it is now over 85% of their portfolio.

So even though the stock market went up a lot, they own a lot of Bitcoin.

It's fascinating to me that...

One of the things that I think is interesting here is the non-deterministic nature of LLMs means that it's very hard to tell whether something changes.

And it's fascinating to me how, you know, like we call it like Opus 4.8, but like the Opus 4.8 that you got at one point in time might be different from what you get at a different point in time.

And, you know, like in any other part of software, of the software industry, like a version... would have a concrete meaning.

We'll leave it to John Boy to go tell you about the fourth out.

The other thing we should update folks on is we got the recording up from last week, and one of the first comments online is like, I, for one, would look forward to LLM-based property disputes, which I thought it was you promptly told Kevin at the Blue Sky not to encourage me, but it was too late.

I was encouraged.

So I think we're clearly going to have to do that as an episode.

detailing all the ways their

LLMs have been used to try to create threats, biological weapons, and on from there.

But we've talked about the opposition data centers not as much perhaps being about the land use or even about water, which is now a very different system than it had been previously, or even power given how much of it is now what we call behind the meter, meaning they're actually generating their own power, but it is simply the physical manifestation of this technology that a lot of people, for good reason, are worried about.

And now stories as well today that Congress seems to be waking up to the concerns of its constituents.

And I feel it's so helpful to really, really demonstrate that and to walk people through the process so that they can really kind of grasp what's happening under the hood.

And you can better kind of set the conversation up to explain the incredible difference from search engines over here, which is true information retrieval to LLMs and how they're kind of working together today.

And that working together is kind of important.

We're going to get more deep into the details, I think.

But I think a lot of people have misconceptions about what's going on here.

Or you pointed it there, you know, they think of LLMs as actually information retrieval as well.

What are the kind of most common misconceptions that you see?

Oh my gosh, that they rank anything.

It's called retrieval augmented generation.

It's a way you go and retrieve information so that when you use the LLM, you can bring really relevant context to it.

It's turned out, I think, to be the perfect choice.

It lets us move between foundation models And today, we support OpenAI, we support Cloud, we support Gemini, and we get good results from all of them.

both have things whose outcomes you can predict.

You have to be able to tell a regulator how you're going to handle the situation.

And just telling the regular, I'm going to let the LLM decide, I think that's pretty unsatisfactory.

And you need to be able to change the way you handle situations as circumstances change.