This a pretty neat idea, but I am wondering whether it could be improved. For instance, if the user’s reply is not satisfactory, the bot should not only explain why, but also let the user to try again. Feel free to share your own, improved prompts.
PS: I am not the author of the Small Talk Simulator.
For more than a year, Microsoft’s engineers and researchers have worked to create personalized AI tools for composing emails and documents by applying OpenAI’s machine-learning models to customers’ private data, said another person with direct knowledge of the plan
Really Nice Images (reallyniceimages.com) are pretty good. They have an iPhone app and also Capture One and Lightroom/Photoshop profiles (which you can use also with Lightroom mobile).
Btw, please don't editorialize titles - that's in the site guidelines: https://news.ycombinator.com/newsguidelines.html ("Please use the original title, unless it is misleading or linkbait; don't editorialize.") If you want to say what you think is important about an article, that's fine, but do it by adding a comment to the thread. Then your view will be on a level playing field with everyone else's: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...
Thanks for pointing this out. I attempted to use the original title, but it was too long to fit within the character limit, so I chose a different one instead. Is there a rule to follow if the title is too long? Should I just let it get cut off?
Sorry for the delay in replying! When titles exceed 80 chars it's best to drop superfluous words until you get something that fits the limit and is still recognizably the original title. In this case the other submitter simply dropped everything after the colon and got a usable title.
If that's not possible, the next best thing is to look for something like a subtitle or subheading, or a phrase from the opening paragraph, that says what the article is about.
The US will see a slowdown in GDP growth, with a rate of 0.2%, while the Eurozone may experience a mild recession, with a rate of -0.1%. Inflation will remain high, particularly in Europe, though it will be lower than in 2022.
90% or more of the population of China will become infected with SARS-CoV-2 Omicron variants during the first quarter. Assuming infection fatality rate of around 0.1%, between 1 and 1.5 million people may die, however the Chinese government will report numbers that are lower by at least an order of magnitude. The supply chain will be impacted, and it is possible that new variants may emerge and spread to the rest of the world.
The Russo-Ukrainian war will continue. A second major offensive against Ukraine is likely. Mobilisation (both military and industrial) will continue, Ukrainian infrastructure and economy will be further damaged. There will be no negotiations. Russian regime won't collapse, Russia will not split, NATO will not become involved in fighting on Ukrainian soil. The risk of a full-scale nuclear war is very low, the use of tactical nuclear weapons is more probable, but still unlikely.
China will not launch an attack on Taiwan in 2023.
There will be many new startups based on the OpenAI API, and this market will become extremely crowded. OpenAI will continue updating their GPT model and release version 4, which will significantly improve the quality and reliability of text output and will be multimodal. There will be at least one major, publicly available, competing LLM, likely from Meta.
Twitter will add the ability to make payments on their platform and will work on developing e-commerce capabilities. Meta will follow suit, most likely on Whatsapp.
While I don't know what IFR to use for China -- I'm not familiar enough with their vaccines nor their rate of vaccination -- remember that in general, there's one IFR for "sick people can go to the hospital" and another, much higher one for "sick people cannot go to the hospital because they are full."
The hospitals will be full, so the IFR will be at the high end of whatever we fear.
You've got a good point there, but we also need to consider the time frame. I'm talking about the first quarter and my predictions may be way too pessimistic. Once the virus infects the urban population (~60%) it will slow down. IHME forecasts 300k deaths in Q1.
The subvariant spreading in China is Omicron BF.7 with basic reproduction number of R0>=10. That means that herd immunity will be reached at 1-1/10 = 90% of population or higher, so by the end of 2023 the death toll may indeed exceed 1 million, depending on the exact IFR which is still uncertain.
