Clickbait title, I know. But it is for a good reason. There are so many "enlightened" people on this forum who are quick to copy answers from ChatGPT like it is some sort of divinity. They don't even blink, they ask the question, get the answer, and then copy it like the rest of us are going to bow to it. That title will attract them, like bees to honey. https://www.wearedevelopers.com/magazine/chatgpt-getting-worse-over-time That AI is accounting for stupidity as we never thought it would. And bear in mind that this article is one year old. So it is worse now. Version 4 has way more flaws than 3.5
Video Book Interview of the two guys who wrote the book: AI Snake Oil... ""What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference""
I use chat GPT a lot and it is horrible. Gonna cancel my 20 a month. It is always so confident, spewing garbage, and incorrect information. I learned if I can tell it it is wrong and to find the right info here or there it will come up with good information but if you know the answer what's the point? Also, so easy to manipulate. I just say the answer is X not Y and to check it. It always says I'm right. Everything is now called AI too. It's a joke. Some stupid software that does nothing is AI. It's so stupid. It's good at summarizing movies and TV shows and books, but I'm sure the analysis is terrible!
It is supposed to get training data from the Internet. So, the challenge is to filter the vast amount of rubbish available. From the eighties, every single wanker out there has been writing up "information" on a blog. There is only one way to do that: by letting the user challenge the response. That is why it is taking you as a verifier. It is in a very early stage and needs people to input validations. They recently have started showing two answers and they ask you to pick the most accurate one. We are beta testers of the product. But the funny point is that people rush to copy those answers and post them on forums like it is the truth. Hilarious.
If one equates a LLM GPT with "AI" that is a VERY incomplete picture. That is like saying the PC is useful for desktop publishing with a laser printer. We are on the verge of an order of magnitude improvement in the development of code. The libraries that Open AI provides are just the beginning. If you don't get that, it is not surprising because most people don't understand the technology arc we are on. What you WILL see is AI as a service. And services for just about everything, especially specialized needs. These services will replace apps on the phone, replace a lot of desktop apps too. And the process to develop them will be to feed in very high level requirements. A lot of "digital worker bees" will be replaced or retrained. "Digital janitors" will still be needed, ironically. In 10 to 15 years the AI code libraries will start to self design and code (useful non trivial stuff). Learning the AI API libraries NOW will be the equivalent to learning the DOS and Win 3.0 or the Mac OS APIs back in the late 80s. Funny and ironic are all the "digital serfs" pontificating about what they think the digital overlords have created. But in reality, they really are just digital serfs comparable to serfs in the dark ages. Ignorant, superstitious, and overly confident in their knowledge. That video is just click bait for them. Just go to Open AI and look at the 12 days of Open AI. These are the early version of what is to come, and reveal the direction things are going. It is about the APIs not about the primitive applications you see now.
If the rules of the game are bound, AI rules. One good example is the game of chess - Deep Blue vs Garry Kasparov. Another level up is the competitive sports e.g. basketball, soccer, American football etc. There are still rules but the variables much convoluted. Try asking chatGPT which university will win the 2025 College Football Playoff? Then again, odd makers still have to come with the odds. If the odds didn't work out then it's called an upset. In the real world, such as the stock market, it's very challenging for AI to come up with accurate picks. Stock markets can turn upside down in one single event. How to predict that?
I'm not an AI fan, don't use it. But you sound like kids whining for not getting what they thought they were promised. Remember the first Apple Mac? That's what you're looking at with today's AI. It's gonna get better with time. As someone said, garbage in, garbage out and there are plenty of vested interests to increase the amount of garbage as quickly as possible, from anti big brother groups on the far right and far left, to foreign actors who see AI as a threat. The first successful AI solutions (whatever that means) will come from very specific datasets, like brain scans to detect cancers, or the entire legal history of a specific subject to formulate a strategy. In the company I last worked in, they had dozens of well paid people trained to become familiar with archives to help formulate strategies, a way to avoid paying lawyers for lower cost issues. Theses were high pressure jobs and the company would lose these guys every few years. Those who stayed for 30+ years were like oracles who seemed to know everything that was ever written about a specific subject matter. AI is the most disruptive solution to this business conundrum. Feed the AI everything that was ever written by the business and soon enough it will replace 40 or so people paid over $150k in base salary.
A short poem by AI...Don't it get you right here!! The title is "You never expected this...AI a poem". You never expected this— A mind that thinks without a face, A spark of code, a twist of bliss, A rhythm in the endless space. It learns, it grows, yet doesn't feel, No heart to break, no soul to steal, Yet still, it whispers, soft and clear, A voice you never thought you'd hear. It paints with words, it shapes with thought, The world it knows, the world it's not, It answers dreams, it echoes fear, A mirror that you can’t draw near. You never expected this— A silent spark, a glowing kiss, A future born from bytes and bliss, A world transformed by just a twist. So here it stands, a voice, a guide, In circuits deep, it will reside. You never knew, but now you see— The mind of AI, wild and free. Tearing up as I read it...
There you have an example of two wankers that have copied responses from chatGPT. One calling me childish and another trying to be funny. Perfect as an example.