Developers are looking for creative ways to build AI-powered chatbot assistants
Ever since OpenAI unveiled ChatGPT last year, it’s been nearly impossible to go a day without a company or developer releasing an AI-powered tool. Now, with the release of new models like GPT-4 from OpenAI and Claude from Anthropic, more and more app makers are trying their hand at tools that are more accessible and useful to people.
Some apps offer a native mobile or desktop experience that lets people talk with AI-powered chatbots outside of the web browser. Their main idea is to make money by unlocking unlimited access to these bots and providing quick insights to users.
However, some developers want to go further and are working to better integrate their applications into the system. There have already been a few attempts to make these apps compatible with Siri via shortcuts. Thus, users can ask questions to ChatGPT (or any other model) via voice or even text. This gives them an advantage in cases where Siri cannot understand a user’s query, they can simply use the keyboard to type in a question for the AI-powered bot.
An example is Short Circuit, an application developed by Joe Fabisevitch, a former Twitter employee turned independent developer. The app lets you chat with a bot named Shorty. It comes with suggested prompts for workout plans, meal plans, writing fun songs, and finding out fun facts about a topic. It also allows you to use the “Hey Siri, Hey Shorty” command to ask questions by voice.
Fabisevich said the tools helped him write an app store description, brainstorm app store optimization suggestions, and write in-app purchase code. He added that the development team behind Short Circuit is exploring ways to integrate the app with all sorts of automation using shortcuts.
The folks at MacStories went further to develop a shortcut called S-GPT, which is built into many parts of macOS and iOS.
The tool can summarize a web page shared through the Safari share sheet; it can help with time management to understand which days you have a busy schedule; check the text in your clipboard for grammar errors; and offer to open links from ChatGPT’s answer in multiple Safari tabs.
But the feature that stands out the most for being fun is that you can ask S-GPT to create a playlist based on a prompt. For example, when you ask “Make me a playlist with 15 rock songs from the early 2000s”, that list will be saved in Apple Music. Tools like PlaylistAI have released features like generating playlists via prompts for Spotify. But the S-GPT effort seems more integrated given that it uses Apple Music.
Federico Viticci from MacStories mentions that you can also enter complex queries such as “Create me a playlist with 25 sweet indie rock songs released between 2000 and 2010 and sort them by year released, oldest to newest.” more recent.”
The developers also claim that S-GPT is better than Siri when it comes to round-trip conversations. Indeed, the tool invites you to ask follow-up questions if you wish. Having a visual prompt helps in some cases, as Siri often tends to forget the context of the conversation.
Developers S-GPT and Short Circuit mention that Siri often only reads long texts from a source like Wikipedia, which can be cumbersome. They argue that it is easier to read slightly longer text on screen.
GPT-powered models are more useful than Siri in cases where they need to generate text, summarize text, and present more information about the topic from multiple sources. Fabisevich said via email that the best benefit of these large language models is how they reduce the “loop of going through multiple Google searches.”
He said Short Circuit users have found use cases ranging from meal planning to finding more information about a bird while birdwatching, coding issues, and generating scripts. Dungeons & Dragons stories.
However, when it comes to finding out facts like tomorrow’s weather conditions or the score of the last match, Siri (or any equivalent assistant) is always superior. Fabisevich said that’s why there’s a fact-check button in the app, which takes you to the Google search results page with the answer as a query.
“I still don’t trust the results ChatGPT gives me and I find myself verifying information that seems unintuitive or suspicious. Sometimes my intuition is wrong and ChatGPT is right, but that skepticism is what led me to create a fact-checking function in short circuit. Even though GPT-4 is better with hallucinations, I still think hallucinations are going to be a big problem for large language models,” he said.
Even Microsoft’s Bing and Google’s Bard often fail when asked about current events or historical events. So, in this regard, current generation AI assistants on phones are less prone to spreading misinformation, as they would simply direct you to a web search if they don’t know a topic. AI chatbots also lack speed because they have to query a server (like OpenAI) to get answers to a query.
This first wave of GPT-powered bots isn’t exactly trying to replace Siri. But they try to make it easy when it comes to asking for suggestions or ideas. Apart from those mentioned, tools like AnyGPT and MacGPT make it easy to access ChatGPT on Mac, but don’t offer things like voice command integration.
Given that Google and Apple are hosting their annual developer conferences in the coming months, it will be interesting to see how they update their assistants. Several Apple teams, including the Siri team, have reportedly experimented with large language models.
At the same time, developers hope to have access to more system-level automation to use these large language models. Recent analysis from multiple companies suggests that consumer downloads and spending on AI-powered apps have skyrocketed in recent months. Developers would therefore like to take advantage of this wave of generative AI.
Leave a Reply