Cortana Design Logo
Twitter LinkedIn RSS Search

Trent Steenholdt's Blog

A simple blog sharing my findings, interests, and work in the Information Technology industry, as well as anything else that interests me.

LLMs and How They Will Play a Big Part in the Startup Game

trentsteenholdt
April 18, 2023

3 minutes to read

Note: This was originally posted at MakerX’s blog at https://blog.makerx.com.au/llms-and-how-theyll-play-a-big-part-in-the-startup-game/

2023 has been a transformative year for artificial intelligence, affecting many industries and sectors. The rapid progress of tools like ChatGPT shows how AI is changing the way we work and think about large language models. Ideas that were once only explored by a small community have now become mainstream. A good reminder of how far things have come is the contrast with Microsoft’s Twitter bot Tay, which struggled with even basic language challenges. In this post we explore LLMs in general and how they are reshaping the startup and venture building space.

Large language models

If you think LLMs are incredibly intelligent you are not wrong, although the real secret is the data behind them.

“Data really powers everything that we do.” Jeff Weiner

LLMs such as ChatGPT work through neural networking and next word prediction across vast datasets. They learn patterns and structure from language and apply these patterns when prompted. If you are new to these concepts, this CGP Grey video remains one of the clearest explanations of how these systems work.

Entering the dawn of autonomous LLMs

Auto GPT at play inside Visual Studio Code
Auto GPT at play inside Visual Studio Code

Autonomous LLM systems such as Auto GPT introduce a major shift. They can gather information from the internet in real time, unlike ChatGPT which is limited to a dataset ending in September 2021. This means they can remain informed about trends and current events.

This development has opened a world of possibilities for anyone with an idea. As one example, Auto GPT can summarise local news each morning, saving time and manual browsing. It is now feasible to prototype ideas, validate concepts and work through many common startup challenges such as go to market strategy and financing.

Startups and LLMs; what is possible now

Based on current interactions with autonomous LLMs, here are some areas where they can already support startups and venture building.

  1. Social media manager
    They can help build social strategies quickly. Tools like Auto GPT will continue to improve, producing images and videos that require little refinement. It is only a matter of time before LLMs generate memes.

  2. Marketing
    ChatGPT is already useful for shaping go to market strategies and challenging assumptions that often slow down or block early stage ideas. With the number of integrations and helper tools appearing, it is clear that any idea can now be explored and iterated on.

  3. Research
    Auto GPT makes it straightforward to research whether an idea is already in the market, gather feedback from existing communities and identify unique positioning options for your own product.

  4. Legal
    Privacy policies and terms are often left until late in the process. OpenAI based tools can now inspect your repository, read code and documentation and generate tailored policies based on what they find.

Auto GPT exploring Canva to understand how to create social media graphics
Auto GPT exploring Canva to understand how to create social media graphics

Conclusion

LLMs are making startup creation and venture building more accessible and less intimidating. With a wider audience now using these tools we will see an increase in innovative projects and new ideas gaining traction. It will be interesting to watch how this technology develops and which teams will be first to unlock its full potential in the startup world.