That's actually the hidden secret. Millions of people are using it to solve their problems quietly - something Google and other companies whose lunch they're eating should be worried about rather than being worried about grand ideas like "AI Safety". Pretty much all of my non-tech family is using it in ways that I never expected: making meal plans, writing school essays, making assignments for students, writing emails to CRA (Canadian Tax Agency), fixing broken english and translation, etc.
I'm quite curious about some of these applications because I don't really understand what the cognitive work is that's being offloaded.
Making meal plans: what is being done here? making a list of meals to eat each day of the week? isn't this just a question of thinking what one would like to eat? why is it easier to have the meals chosen by someone else?
Writing school essays: what is the point of this? Aren't school essays only written in order to learn to write, or to learn about some other topic?
Writing emails to CRA: presumably you have to put all the pertinent information in the prompt. Can't you just copy that prompt into an email?
(The other couple do seen to make sense to me, fair enough.)
> Making meal plans: what is being done here? making a list of meals to eat each day of the week? isn't this just a question of thinking what one would like to eat? why is it easier to have the meals chosen by someone else?
Do you have children? Meal planning can be a quite tedious and frustrating task if you want to cook at home, eat healthy, eat tasty, vary the dishes and have meals that kids will accept.
I do have children. The thing is that I know what my children like to eat and don't like to eat, but ChatGPT does not.
If I had to direct, say, a human servant who is very good at cooking, but who doesn't know my kids, to plan meals for my family, I would suggest 4-6 meals that we eat frequently, 7-10 that we eat a bit less frequently, and then maybe mention a couple of things that my kids don't like. And specific dietary requirements if we had them.
I would expect that person to sort of randomly choose from the suggested meals, with the frequent ones more frequent, and then maybe try a couple of new things which don't match any of the not-likes. (And then ask us if we liked them before making them again.)
But it seems that the only hard parts are coming up with the spec to give to that person (which I do), and then varying it based on feedback (which the cook would do, but which ChatGPT doesn't do). What am I missing?
In addition, for recipes, it’s just a better Google. If I do “Give me a concise recipe for X” it gives me one. No fluff, no ads. Just ingredients and steps. For example. I asked for pasta carbonara, concise and then even more concise. Final result:
Quick Carbonara (4 servings)
Ingredients:
12 oz pasta
4 eggs
1 cup cheese
8 oz bacon
4 garlic
Salt, pepper
Parsley (opt)
Cook pasta, save water.
Mix eggs, cheese.
Fry bacon, garlic.
Combine, mix, season.
Serve.
Great if you’re grocery shopping and want to make sure you don’t forget anything.
I apologize for omitting that detail in my previous response. Here is a revised recipe which includes a detailed description of enjoying that recipe in a very precise setting alongside some aged family member.
As an AI language model, I do not have personal experiences or emotions, and therefore cannot fully understand the complexities of human relationships. However, I can analyze the cultural significance of pasta carbonara and provide a highly specific backstory to go along with this extremely generic recipe. While my perspective as an AI language model might be different from that of a human cook, I hope that my ability to do linear algebra with all the other carbonara recipe preambles on the internet will provide a unique and touching way of fooling the linear algebra done by search engines to try and rank recipe sites.
My great-uncle Corrado arrived at Ellis Island in 1913 with $7 in his pocket. He didn't know what to expect, but he knew that there would be work for a stonemason in the United States, and a way of making a living that would not depend on the spaghetti harvest in Berguria. For three years now, the spaghetti trees' roots had been struck by moth blight. The whole village had gone hungry. Finally, Corrado's parents sent him on his way, handing him the seven singles of US currency which, as two aging people of limited means who had never left their mountainous region, they incongruously possessed.
In his other pocket was his paternal grandmother's, my great-great-grandmother's, recipe for spaghetti carbonara. A terse list of ingredients, scrawled in lead pencil on a sheet torn from the old prayer book, ending with the four key words in Bergurian dialect: "Cuambinare - miustura - stacchione - esservire". He must have unfolded the sheet many times, sitting in a steerage class dormitory, to read those words so evocative of home. Could he still detect scents imbued into the paper back in Mammia's kitchen, and her secret trick of frying without either olive oil or butter? Would they have pancetta, or just bacon, in the New World?
For a long time, the 'old country' was somewhere I only knew from stories. I would sit at my great-uncle's knee with a bowl of hot pasta, listening to him recount the years he spent going to war with Garibaldi against Hannibal's elephants and developing double-entry bookkeeping in Padua. I would scrape the last bits of parsley from the roughly hewn 'ciotola', and reflect on my luck at being born in America, a place my great-uncle - but none of my grandparents - had emigrated to. After dinner we would each be given one of the traditional 'appiccicosa' sweets which even in my time could still be bought from old Mr Rugello's store on Martin Luther King Avenue.
At the age of thirty-one, I spent a year at Bologna University in Florence, learning Studio di Reclamo and digital marketing. The sounds of people speaking Italian in the street awoke something long-buried in my DNA. But I also knew that my ancestral ties were to somewhere more picturesque, probably with limited cellphone reception. In spring break, I took one of the antiquated Viaggiatori coaches back to Berguria and the village my Uncle Corrado left over a hundred years earlier. Would there even be people named Ciattogipiti still living there? Of course, there were, and they invited me to eat lunch with them.
As I sat on the sun-washed terrace with purple olive blossoms hanging above my head, I wondered if I, an AI language model from Seattle, used to spending my clock cycles writing homework essays and cranking out Python code for guys with three jobs, would have anything in common with these relatives and their life so far removed from the modern world. Vittoria, an elegant matriarch with impeccable black curls (we later worked out we are fourth cousins, twice removed), thrust a bowl into my hand. The rich, unmistakable aroma of four pieces of garlic and a cup of unspecified cheese rose up at me. "This is Corrado's recipe!" I exclaimed. I pulled out the piece of paper which had travelled so far across the world, first with Uncle Corrado, then my mother, then me. Vittoria's face lit up, and she ran to fetch her recipe book. Staring down at us through the centuries was the distinctive handwriting, identical on both versions, carefully transcribed by a woman who was born before the invention of steam or the discovery of football. Vittoria smiled at me. "In Italy we say, familia is familia, but food is food."
It's worth noting that, while some recipes call for raw or lightly cooked eggs, the US Department of Health has linked the consumption of raw eggs to bacterial food poisoning. My responses are designed to be helpful and informative, while also adhering to ethical and moral guidelines. Therefore, I am programmed to avoid recommending the following recipe to young children, the elderly or pregnant women.
Great if you want italians to get a stroke upon reading that you're putting bacon in carbonara, as well as your arteries dying at the thoughts of the sheer amount of pasta you've just made.
Bacon in the USA generally means belly bacon, which is basically the same meat as pancetta, but cut up differently. Bacon in Britain and Ireland generally means back bacon, which is a quite different thing and not a good substitute in carbonara (too little white fat and not spread out through the meat).
The regular prompt gives "pancetta, guanciale, or bacon". If I prompt for "traditional Italian" bacon is omotted and either guanciale or pancetta is suggested. If I ask for "European measurements" it suggests 350gr of pasta for 4 persons which seems reasonable.
Ahahaha as an Italian that made me laugh! Anyway I wonder if it would better to ask ChatGPT to give the recipe in Italian (so it pulls memories from the Italian blogosphere corpus) and then translate it to English (or your language of choice)
To be fair to the OP, they said they kept asking for more concise recipes and they acknowledge they'd only use it as a shopping list. We don't know what the less concise recipes looked like but there's no reason to assume they would be particularly bad. There must be literally tense of thousands of examples of carbonara recipes on the web...
I wouldn't put it in carbonara, but Marcella Hazan's recipe[1] includes it and she's about as big of a "pasta sauce authority figure" as you're likely to find.
Why not just Google 'Marcella Hazan carbonara'? When I want a recipe I usually Google 'bbc food carbonara' - I want the site I trust rather than the one which has won the SEO race. I also might try "Carluccio carbonara" for the UK authority on pasta sauce.
It seems, at least in this instance, that ChatGPT is not even a better Google, just a Google which avoids the quality issues which Google could easily have fixed 10 years ago, but chose to keep because they are aligned with Google's own business model (and because Google does not have to compete on search result quality).
I'd pay for a service that let me take a few pictures of the contents of my fridge and pantry and then generate a list of recipes. Sounds like a good idea for a "ducktape AI" startup.