What is your AI provider of choice when working with Elements?

Elements
, ,
22

You’ll note that most of the doomsayers are not telling you how AI is going to kill you. This is typical fear-mongering biased by decades of dystopian media (even Star Wars is dystopian). Historically, every new technology has had its dark side; humans can, have, and will use new tech in ways that create problems for others.

What I think most of those fearing the Skynet apocalypse are missing is that it’s going to take a human (sentient being) to deploy AI that kills, just as it was with nuclear weapons and all the previous tech pushes. We need to worry much more about who’s in charge and what their motives are than the actual technology involved.

My fear about AI as it is currently done is that it will result in “stale humans.” AI tokenizes everything we know/create/publish, but if it’s doing a thorough job of that, imagining, creating, and expressing new things goes away as a human trait. The danger of AI in schools is not that it can’t help you learn (it can), but rather that it introduces an incentive not to learn.

23

There’s a reason the C-Suite tech bros don’t allow their own children to use these technologies they are lording over…

24

Since I started with punch cards and CDC6600 assembly language, I’ve done one heck of a lot of learning over the years, some of which resulted in seminal programming books. AI helps me learn faster, though I do have to make sure that I verify (e.g. I use two different LLMs to cross check each other). The problem I see is this is not how AI is being approached in education, even up into the University level. To me, AI is having a really smart library on my beck and call, plus an assistant who can do 90% of the work for me. I’m not sure if that’s the future, but I’m not going to stop learning.

25

In UK there is a hosepipe ban and people keep having to collect bottled water as taps run dry. In Kent there us a brand new Data Centre when it was built the water authority sent a letter to all residents 144K of them to say their water would be restricted and there may be shortages due to the data centre. It was asked to use water from a ttreatment centre but they refused. Therefore AI is killing the locals and ruining their water supply now. Waiting for them is not an option!

26

(e.g. I use two different LLMs to cross-check each other)

I do the same. It started because one model was better at certain things than the other, and I needed the strengths of both. I would send a piece to one for polishing or a bug fix, then give it to the other. The second model would often find “issues” the first had missed.

That back-and-forth became a kind of checks-and-balances system that produced a better final product. The bigger benefit came when I forced myself to read the errors, understand the corrections, and figure out the “why” behind them. That is when the results became almost magical.

I think Thom is onto something.

Recently, I was auditing some financial records. AI was excellent at drilling into the numbers, finding technical issues, and spotting patterns. What it struggled with was interpreting those patterns while keeping track of the large number of variances, explanations, and additional data that an audit requires.

I had to keep it on track and repeatedly ask it to double-check calculations. I had to teach it the sources, the whys, the whens, and the hows behind the numbers.

If I had never learned the fundamentals of the job myself, I could easily have accepted the AI’s initial conclusions.

And the difference was enormous.

AI initially saw a failing business with signs that raised suspicion of fraud. After the records were fully examined and the patterns were put into their proper context, the picture changed completely. What we actually had was a very profitable opportunity with a legitimate, verifiable explanation for the unusual numbers. For my client, this particular business acquisition turned out to be a very good deal.

AI was extremely useful in finding the pieces. Knowing the fundamentals plus my real life experience with the industry this business was saved us probably a million bucks over the next 5 years.

27

I think the authorities are culpable for this. They make the rules. AI is just a machine and data, it did not “take” the water. Some human authorities really messed up if this happened.

28

I presume you mean hosepipe ban? :wink:

29

Nope. AI is not doing the killing. Your local politicians who are allowing that are.

30

dont use it…it will ruin the planet eventually

31

Claude Code for writing and managing my agency, and ChatGPT for images.

32

Official Elements docs talk about using your own API key and specifically mention OpenAI and Anthropic in their existing writing-assistant material. I did not see xAI / Grok listed as a first-class provider in the current Elements AI pages. So Grok’s API exists, but Elements may not yet have a “paste an xAI key” option when the upgrade ships. That is something to confirm in Elements Settings → AI after release. Will xAI be a provider?

33

We’re not sure how popular the built-in assistant using your own API keys will be, so we didn’t want to support every single provider just yet.

However, you’ll be able to use most providers via the Elements MCP server, including Grok. Does that help?

34

hal
hal1333×1000 45.3 KB

… HAL obviously . . . then Gemini / ChatGPT, no subscription.

35

Understood and appreciate the option. thank you

36

Using Claude Max for working with code and SEO, and ChatGPT Business for marketing, writing articles, emails, social posts, etc… I used ChatGPT today to build out some features to our Elements CMS template and watched it open RW Elements, move a mouse, click buttons, write the code, etc… it was fun to watch, but I kept reviewing the pages before it was completely finished and then prompted me, ‘I’ll stay out of the RW Elements project until ChatGPT finishes’ Click to Accept. :joy:

37

Just like everything else, I think it will do as much harm as it does good.

38

I might be - too optimistic - but I think it could be - wonderful as AI could fix many problems - and help people who are literally passing away from malnutrition and lack of clean water….

So I am hopeful for the future!

39

AI = laziness

think of kids that cant do basic maths now cos the invention of a calculator…which are allowed in exams these days. what happened to show how you got the result.

Anthropic in the news yesterday cos people were using it to research weapons of mass destruction.

AI would be useful if it could stop spam f’in phone calls!!! but it cant even do that…Im guessing Jason Arday ended up where he did partly because of AI..had to use AI to write his resignation letter…I can imagine the future…surgeon why did you cut that bit when the patient died…cos Claude told me to!!!

40

so how will AI achieve that…create money?..create politicians that know what thier doing?..we have labels on car batteries to not drink the acid…we have politicians who think women can have a penis!!! and your hopeful for the future!!! good luck with that!!!

41

Some examples - scientific discoveries will change the world for the better - power: unlimited- clean etc…

  1. Disease cures — protein folding and drug discovery at superhuman speed.
  2. Pandemic detection — spotting outbreaks weeks before humans notice patterns.
  3. Climate modeling and materials — better batteries, fusion control, carbon capture design.
  4. Food security — crop optimization, drought-resistant breeding, precision farming.
  5. Scientific acceleration — compressing decades of research into years across every field.
  6. Medical access — expert-level diagnosis for billions with no doctor nearby.
  7. Disaster prediction — earthquakes, floods, famines forecast early enough to evacuate.
  8. Asteroid and threat monitoring — continuous scanning no human team could sustain.
  9. Education at scale — a patient tutor for every child, in any language.
  10. Coordination tools — translating, modeling negotiations, finding deals humans miss.