As the large language models took the Artificial intelligence domain by storm, all the science fiction movies a whole generation grew up watching, people speculating about an era ruled by robots, all the heated debates about the ethics perspective. All of these points are pushed further into reality as the capability of ChatGPT and now GPT4 stunned the whole world.
While users were still playing around with the chatGPT3, exploring the tip of the iceberg, OpenAI revealed GPT4 LLM, which can not only “read” the texts you input but can also “See” and “Understand” the images you give it to. Better trained, better fed in terms of data, and manifolds better in terms of processing the input data in one go, this marked the beginning of the race of building more AI-powered tools that can make a dent in the computation universe we are aware of as such.
Such a tool is Pico’s MetaGPT. This powerful no-code solution to app building knocks all the expectations out of the park.
Simply put, this web service allows you to build other web apps without any prerequisites of any coding language or, as a matter of fact, any technical experience.
Whatever you can imagine, you can build. It simply asks you what you want to develop, then asks for the prompts you wish to from the user, and then based on the prompt, what you want to do with it, and what actions you want to take. Once fed with this information, it can spawn a web app for the same. Pico claims that it can understand all the major languages in the world, so language locality as a barrier to interaction with the app is gone for good. It spawns the app for which they provide hosting and a shareable URL to let others try what you have built. We tried it out here, and it seems very much on point.
MetaGPT is highly customizable, allowing you to reiterate any changes you wish to incorporate. These iterations can be anything, ranging from a design change to a bug fix.
Users love this solution so far. The project is not only fascinating, but Pico’s MetaGPT is also solving a crucial problem: fueling the usage and the research driven by the proof of concept notion with the interaction of these LLMs as the world is aware of OpenAI’s GPT plugin market as an attempt to build an extension market powered by GPT engines, a business model quite similar to the play store or apple store that we all use regularly. Although this seems to be quite a lucrative solution, currently, due to the hype of these models, there is a long waitlist and very limited access to any general user, which is kind of a setback to ML/AI tech enthusiasts that want to poke the model to understand its functioning better.
Developer community-driven projects that picked onto OpenAI’s slack to learn the deployment of GPT-powered APIs in a better manner has created some amazing projects such as AutoGPT, which have accelerated the craze and shifted the attention from GPT LLMs to such projects which incorporate these LLMs in a better and useful way.
The delay in the launch of OpenAI’s GPT4 multimodality feature is creating incrementally increasing space in this domain, which adds to the research and development of such projects, enhancing the applications built on current capabilities and widening the horizon by introducing new use cases.
With GPT engines into play, communities involved in rapid development, and a hunger to be better, to “understand” and interpret a “use case” more efficiently, the line between artificial reality and actual reality is thinning exponentially. Widespread awareness of such technologies, and the buzz it creates, does raise certain concrete concerns over the usage of such advanced models, but at the same time instills an age-old human nature, curiosity. Curiosity to know how much further we can go and how much more we can achieve.
Check out the Tool Here. Don’t forget to join our 20k+ ML SubReddit, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more. If you have any questions regarding the above article or if we missed anything, feel free to email us at Asif@marktechpost.com
Anant is a Computer science engineer currently working as a data scientist with experience in Finance and AI products as a service. He is keen to build AI-powered solutions that create better data points and solve daily life problems in an impactful and efficient way.