Notes from a CTO #7: USA here we come, User experience, GPT in production
We are in USA to understand client's pain points and build a killer product. On LLM we have finish first phase experiment and few plan in this quarter.
Hey, it’s Bkrm from Docsumo.
You’re reading the seventh issue of Notes from a CTO, my raw canvas of thoughts and collection of interesting resources I found online. My goal here is to spread knowledge on a large scale. If this is up your alley, have a read and let me know what you think!
1. Running a startup
This week, let's discuss our plan for the USA.
To understand client's pain points and build a killer product, you need to be closer to them. This is the advice we have been hearing since we started Docsumo, but it took us four years to finally be here. The main reason we came to the USA is to meet our customers in person, observe how they use our software firsthand, and have clients in every building in New York City.
Moving to a new country is like starting your life all over again. From obtaining a Social Security Number, setting up a bank account, getting a driver's license, apartment hunting (14 done till now), and getting familiar with the culture (paying $6 for the airport trolley), it has been quite a ride in the US in the last two weeks.
While we loved the weather and greenery in California, we will be on the East Coast as most of our customers are located there. We are planning to start our day at 6 a.m. to have an overlap with our team in Nepal and India.
Meeting customers has been the best experience so far. From hearing their stories, how they were used, a planned visit to see how they were used, and the impact we were making. Our goal is to have clients in each building in downtown New York than all over USA and the world. This is one more step in that direction.
2. Technology
Product and user experience are crucial factors for the success of any product or service, regardless of the hype surrounding new technologies or trends. While emerging technologies can bring exciting new possibilities, they should always be used to enhance the user experience and the value proposition of a product, rather than being the sole focus.
Web 2.0 is an example of a technology trend that had a significant impact on the way we use the internet, leading to the emergence of new business models and platforms such as social media, online marketplaces, and cloud computing. However, the success of these platforms is not solely due to the technology, but rather to the way they were designed to address user needs and provide a seamless user experience.
While some technological hypes may have a more lasting impact than others, it's essential for companies to prioritize product and user experience as they navigate the ever-changing technological landscape. At Docsumo, we have started thinking about things from the ground up moving forward. I have always been more concerned with technology than with product and user experience, until now. Once you stop wearing a technology hat, your thinking changes drastically.
Meeting customers face-to-face and attending conferences in the United States over the last two weeks has had a significant impact on my thinking process. This week, I went through the product and noted each and every step. Now that we have completed one round of detailed testing on the LLM model (chatGPT, GPT-4, and others), we are brainstorming and finalizing our OKRs for this quarter. Most of these objectives are related to improving the user experience and making sure people get a WOW moment in less than 30 minutes of using the product.
Here is an interesting take on the current trend of generative AI.
I agree with this thought. One more point to add is that distribution and user experience will separate the winners from the losers in the race to the top. And this is true for any product.
3. Podcasts/Essays
I didn’t have time to listen much, but I have been reading a lot lately.
Here is a long article I recommend to anyone interested in or working in the AI space.
Building LLM applications for production
This article cuts through the hype and provides valuable insights and reflections on the fields where AI might have a big impact and what is still left to be done.
After the first phase of experimentation on our side, we also came to the same conclusion that the new LLM will have a significant impact. However, we should focus on its strengths that have the most impact on the customer experience, improve the product, and integrate those improvements this quarter. We'll keep an eye on new developments as well, like AutoGPT.
4. Interesting links
Repos:
the-algorithm: Talk of the town, Blog. Here are some interesting stats: Half of Twitter Blue subscribers have less than 1,000 followers, which means retention is going to be a big issue.
Hocus: I am a big fan of CodeSpaces and GitPods. I would love to see some good open-source projects that provide ready-to-code disposable development environments.
Interesting GPT development: Fastchat, AutoGPT, awesome-chatbot-prompts, Lex-gpt, Prompt-Engineering-Guide, AgentGPT. AutoGPT and AgentGPT look interesting.
For more follow me on Github: bkrmdahal
Articles:
Not even wrong: predicting tech: Predication next thing is hard. Dot doesn’t connect backward not forward. One of the things that can help you predict that is if there was a very clear and obvious path(roadmap) to make it better. But this can be tricky like in the case of VR or self-driving cars or maglev. Is there a silver bullet? No, we just need to be better based on our experience.
K3s Vs K8s: I like k3s for various reasons, so it's good to know when to use what.
Adobe Firefly Vs MidJourney V5: Firefly is only trained on Adobe Stock and fully licensed images. The data curation is very conservative, which may cripple its performance. Still result are interesting.
Remember Groupon?: From $16B to $100m. Easy to comment once it’s done but I keep on thinking about what was going on founders’ minds.
Making Python 100x faster with less than 100 lines of Rust: big fan of using tools that are the right fit and are used based on all the words. I recently read a nicely written blog on how to debug performance and also how to write just the part that needs to be written in Rust.
5. Quotes/ Books
But the fact that some geniuses were laughed at does not imply that all who are laughed at are geniuses. They laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown.
- Carl Sagan, Broca's Brain: Reflections on the Romance of Science
There is a thin line between being smart and being crazy. When building a business, this line is even more difficult to navigate as many factors come into play. It's not just about having a good idea but also being at the right place and time, your execution, your team, luck, and more. It's easy to look back at Steve Jobs and say he was a genius in retrospect, but there were many ups and downs in his career, and not everything he did was successful. That is also the reason why some very successful entrepreneurs fail at their second startup.
Our Slack channels have a great collection of memes.
That’s it for this edition. I hope you find it useful.
P.S. I hope to resume my schedule once I am set up here and start my biweekly sprint of these newsletters.
Best,
Bikram Dahal
P.S. If you learned something new today, please share “Notes from a CTO” with your friends and spread the love. ✌🏻