• Scam Alert. Members are reminded to NOT send money to buy anything. Don't buy things remote and have it shipped - go get it yourself, pay in person, and take your equipment with you. Scammers have burned people on this forum. Urgency, secrecy, excuses, selling for friend, newish members, FUD, are RED FLAGS. A video conference call is not adequate assurance. Face to face interactions are required. Please report suspicions to the forum admins. Stay Safe - anyone can get scammed.
  • Several Regions have held meetups already, but others are being planned or are evaluating the interest. The Ontario GTA West area meetup is planned for Saturday April 26th at Greasemonkeys shop in Aylmer Ontario. If you are interested and haven’t signed up yet, click here! Arbutus has also explored interest in a Fraser Valley meetup but it seems members either missed his thread or had other plans. Let him know if you are interested in a meetup later in the year by posting here! Slowpoke is trying to pull together an Ottawa area meetup later this summer. No date has been selected yet, so let him know if you are interested here! We are not aware of any other meetups being planned this year. If you are interested in doing something in your area, let everyone know and make it happen! Meetups are a great way to make new machining friends and get hands on help in your area. Don’t be shy, sign up and come, or plan your own meetup!

Anyone using ChatGTP?

Mcgyver

Ultra Member
Wowzaaa! (that would be ChatGPT if a mod can edit the title)

This stuff is neat. So incredible that you have to keep reminding yourself it's not perfect. I've a few electronics projects its working on (takes up to 10 days for all the deliverables). Reasonably complicated and if you type a good scope of work and operation criteria it does everything - recommends components, makes suggestions etc then produces a BOM, schematic, software, enclosure design, gerber files, spice files etc. Haven't got the first one back so well see .... but based on well it understood what I wanted I'm optimistic.

Solving problems its amazing. You know when something needsfixing, say on the computer, and the cause is not one the top most common causes? its nearly impossible to search through that because every search response addresses the common reasons. Not with Chat.... keeps the context of the conversation .... you just tell none of those worked and it's digs deeper and interacts. That alone has taken so much frustration out of my days

Trying to give the wife an idea of it, I said lets ask it to create some random image and see what we get.... A barn, five Scottish longhorns at sunset and make it like a water colour:

First pass, I tell it "why are they staring at me, and it looks like six cattle"


ChatGPT Image May 30, 2025, 07_08_41 PM.png


The it produces this which is cool, but little oddities occur - like why did it take the door out? No idea

ChatGPT Image May 30, 2025, 07_12_10 PM.png


Next I say, looks good, how about an oil painting done with a pallet knife - probably not any fancier than what photoshop can do, but I think its neat because it conjured the whole thing. And the door is back in!



ChatGPT Image May 30, 2025, 07_14_15 PM.png


Next how about a high res black and white image as realistic as possible. Notice how its created depth of field

ChatGPT Image May 30, 2025, 07_31_29 PM.png



Pretty neat and you own the copyright .... more than good enough if you need such an image for an ad.

Then I asked for a winter scene - love how it removed the foliage and changed the DOF so you can see the branches.



ChatGPT Image May 30, 2025, 07_39_10 PM.png


Let's make it stormy, and more natural around where they've been eating. It's not cut and pasting, there subtle difference in the barn geometry and position of the cows heads - look at the head position of the little guy second from right for example

ChatGPT Image May 31, 2025, 08_03_52 AM.png



Friggin amazing imo.

Next day I ask for the same thing and get this ..... it still likes the dominant Bull staring you down, but the rest has changed so its not just a library of stock photos. Rendition to rendition, subtle things change like the tilt of the cows heads, trees, barn geometry, so i think it is creating an image each time.


ChatGPT Image May 31, 2025, 11_36_32 AM.png


cool, but lets the get the ground where they're eating and walking more realistic, plus lighter and more contrast and what the heck, make it snow.

Anyway, thought you might like to see this little excercise if you hadn't played with it yet. Amazing possibilities ..... and a bit scary the enormity of change it could bring about


ChatGPT Image May 31, 2025, 11_43_07 AM.png
 
Last edited:
Ya, I've been loving it too. I've mentioned that on the forum before. You must have missed it. I don't particularly like ChatGPT. I like Perplexity and Gemini. Making paintings and cartoons are only one advantage. I agree that it starts somewhat fresh every time. The minor changes you found bug the rat shit out of me. All I wanted was a slight change. I did find that you could sometimes ask it to keep the previous suggestion and build on that. Otherwise you don't know what it will come up with.

I tried making some circuitry too. It did a good job but not at all as good as @slow-poke's skill set. Still good enough for most regular needs I think.

Where it absolutely changed my world is researching academic papers. It takes a lot of time to read papers and understand them. But AI can read a whole global library of papers in many different languages and give you an amazing summary complete with references.

The other thing I've enjoyed is arguing with it. Most people can't handle debate without taking offense. AI could care less how blunt I am. More importantly, it quickly concedes when it discovers that it is wrong and will develop a position that embraces your arguments and includes its own logic. This ability borders on spooky. It actually gives me the false sense that a real expert has weighed in. I love it.

I do wish the companies that are promoting it didn't have to include a range of forbidden topics. I get why, but I still wish I didn't hit barriers like that.

Another dislike of mine is the potential for mischief. The filthy rich can use it to their own advantage while the common folk are stuck in an old paradigm. I guess it was ever thus though.
 
Yes i occasionally use it both for entertainment and information purposes. I've had it do some calculations for me in the past and recieved wrong answers more than once. I told it was wrong and it agreed and recalculated. Sometimes it got it right but not 100% of the time.
 
I've watched enough movies to know this is a bad idea. The whole AI thing will lead to the book 1984 becoming reality, if it isn't already here.

How do you know if you're watching news, AI or actual real videos now?

I share the concern, but feel like we're already there in part. To huge swaths of the population the truth doesn't matter anymore. Political constituents are fanboys instead of critics - look at what's happened in the US and I feel it here now. In a way it is just a continuation of the internet. Back in the old days you had to earn credibility through education and time in to be an expert, get taken seriously by publishers, vetted by editors etc. Nowadays anyone can publish and continuous learning/research seems a lot like people just finding someone "published" be it a video, blog, whatever who agrees with their uneducated views and biases.

I don't know the answer. And it's in its infancy. Imagine ten years of refinement and polishing? You won't know if I'm real or put here to sprinkle promotions of Bambu and Makita lol

Curious to see what it comes back with on the electronics. I would also doubt its going to give an expert like slow poke a run for his money ..... but I'll never have the time to be sd educated, experienced and stay current that way a pro can. I want a timer, I think 555, i want an op amp I think 741. Ancient I know, but I can't put the time into being current for a project once every three years. Chat GPT changes that
 
Last edited:
I avoid all AI as much as possible because of its insane energy consumption.

Ya, but consider that around 20% of the energy a human uses is consumed by the brain which is only 2% of the body's weight, I think it only makes sense that AI consumes insane amounts too.

IMG_0485 (1).gif


Just kidding!!!
 
I find these AI systems to be immensely useful. I had been using ChatGPT for research (excellent) and coding (not so great) and for "what-if?" questions. More recently I changed to Gemini Pro after wasting a couple of weeks with the chatGPT coding tool CodePilot.

Even the latest versions of these LLMs have such a small context window that the initial instructions and data are forgotten quite quickly, which makes for a very confusing set of responses. The training for chatGPT means that it is aware of world events up until 2022 or 2023. That is quite limiting if current events and technologies are part of your research.

I have to disagree with the discussions about circuit design and the so called schematics produced, which are simplistic and miss the subtleties in the design, for example unless I ask specifically for bypass capacitors or isolation, it simply does not think that deeply. When asked a specific questions about the components required for a circuit, the libraries it is researching are out of date - for example on a question about low-loss FETs, chatGPT recommended the IR540N as an efficient low Rds device and was unable to recommend a suitable modern component.

Fortunately I know enough about enough stuff that I can spot nonsense very quickly, and steer the AI conversation by challenging it's results and assumptions. It is a very dangerous tool in the wrong hands for sure, since we humans seem to believe what is written (or drawn), especially if it was someone else that published the garbage.
 
Haven't used ChatGTP yet and not overly concerned about the future with AI.

One of my favorite podcasts is the Lex Fridman podcast. I enjoy listening to really smart people often talk about the fundamentals of human understanding with uncertainty.

Why be concerned about AI undermining our knowledge of the world when none of us can say with any degree of certainty it is accurate to begin with?

I am interested in trying ChatGTP but it will likely be a winter project.

@Susquatch
You can argue with it??? I will have to introduce it to my wife.
 
Sometimes I’d like it to clone my voice, continue the conversation with my Wife and then send me the executive summary of the things I apologized for and their ramifications.
I had to get hearing aids due loss caused by years of shooting. I never liked them but they did have a side benefit. The wife would say put your hearing aids in she never said they had to have batteries. Expensive ear plugs.
 
We use multiple AI at work currently. gemini, chatgpt the most. It is actually getting a bit scary. unless you can touch someone or see them directly there will be no truth in video.
 
Ya, but consider that around 20% of the energy a human uses is consumed by the brain which is only 2% of the body's weight, I think it only makes sense that AI consumes insane amounts too.

View attachment 65302

Just kidding!!!
Here's what Copilot, Microsoft's version of AI built into its browser, said in reply to Susquatch.

This comment appears to be a lighthearted remark made by a forum user named Susquatch in response to concerns about AI's high energy consumption. The discussion on the forum revolves around AI’s capabilities, its impact, and concerns about its environmental footprint.

Susquatch is making a tongue-in-cheek comparison between AI's energy demands and the human brain's significant energy consumption relative to its size. The idea is that since the human brain—a relatively small part of the body—uses a disproportionate amount of energy, it’s not surprising that AI systems, which perform complex computations, also require substantial energy. Their final remark, “Just kidding!!!,” suggests they’re not making a serious argument but rather playfully engaging in the conversation.

The broader discussion touches on AI’s usefulness, its limitations in coding and circuit design, and concerns about misinformation, privacy, and dystopian implications. There’s also a mention of the environmental impact of AI, with a link to an MIT News article exploring sustainability concerns.

Would you like a deeper explanation of AI’s energy consumption, or are you more interested in the general forum discussion?
 
I've never really found any of these LLM things very useful, they are slower (after all the refinement that is needed) and less precise than I can just do out of my head for tasks they might be useful for. I can type as fast as most people can talk and form my thoughts quite quickly, so I can essentially just straight write what I want faster. Even for code, I get it for new/casual developers where they might not know the pattern or language constructs needed to do what they want, but when you know exactly what you want it's easier to just write it directly.

The image generation stuff isn't there yet, it all feels/looks fake still. Even those longhorns at the start of this thread, they're all not quite right. The only time I've tried to use it is with presentations, but it rarely gets me what I want with the right professional polish on it.

The other thing to note is that this AI isn't anything we haven't had the knowledge of for the last couple decades. The only reason we're seeing this now is because computing power has allowed us to brute force our way into something that seems mildly "intelligent". It is neat but it's not a huge leap forward like a lot of people are making it out to be. All the "AI powered" crap is getting a bit out of hand and is getting slapped on literally everything.
 
At this point in my life, I already think that the so-called News, is mainly only fit to line a parrot cage.

All this AI generated (but otherwise unmarked/unidentified) content can paint whatever picture (literally) the user wants.

I have no fondness for Government oversight, but would be pleased to support any motion that required AI generated content to carry a label saying just that. So long as it does not get confused with real world, real issues, or real (so-called) News.
 
Back
Top