This week, the Apsy team converged in San Francisco from various global locations to attend the annual Techstars Foundercon conference. Amidst numerous impressive Techstars, Apsy was chosen to present a pitch, which garnered a warm reception.
During the event, my interactions with investors and fellow founders revealed a profound intrigue regarding Apsy's capabilities. These insights seem ripe for sharing in a short newsletter.
The widespread success of platforms like ChatGPT and Bard may lead many to believe that AI can seamlessly comprehend and execute tasks that don't necessitate specialized hardware. However, this isn't quite the reality. While it's tempting to think of AI as an entity that matches human understanding of intricate concepts, the truth is that, like humans, AI needs to be trained with specialized skills to undertake specific tasks. Furthermore, guiding users toward their true objectives is another challenge altogether.
To dive deeper into the former, an AI app builder like Apsy employs Natural Language Processing (NLP) to gauge user intent. Yet, this is just scratching the surface. The multifaceted skills a sophisticated system needs to craft an app include recognizing missing data, applying provided information aptly, detecting contradictory requests, and navigating the nuances of app store publication, to name a few.
Addressing the latter point, many users often possess only a surface-level comprehension of their vision. When asked about finer details, they might draw a blank. An apt acronym for this scenario is "TDKWTDK" (they don't know what they don't know). Just as an expert developer or product manager suggests additional features to meet an app's technical or business objectives, an AI should also be able to offer these insights.
To synthesize, it becomes evident why a "generally intelligent agent" like ChatGPT may not be adequately prepared to undertake tasks necessitating a unique skill set.
great
awsome