Notes from a CTO #8: Unpredictable AI Race, Startup Playbook, Own Your Accountability
My thought on current AI Race, best startup playbook that I read frequently and about owning your accountability to have peaceful life.
Hey, it’s Bkrm from Docsumo.
You’re reading the eighth 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. Thought of the Week
This week, let's discuss what is going on in my head regarding the AI race.
It's just getting started, and it's hard to predict who the winner will be. Many questions and opinions are floating around the world. For example, some suggest that OpenAI does not have a moat, while others claim that it does. We don't know whether companies like Microsoft, which have co-pilot on everything and consider AI to be just a feature of their products, will come out on top. Or will it be companies that build fundamental models or companies founded by AI experts who have raised over $100+ million? Or perhaps it will be companies that are bootstrapped and built on top of open source or proprietary APIs, or those that pivot on all this hype and crush it. Will this be a winner-takes-all market, or will different models specialize in certain fields? Is this hype similar to that of fintech (“18-”19), Edtech (mostly India “19-”20), web3 and blockchain (“20-”21)? These are many questions that will be easy to answer in hindsight, but extremely difficult to predict moving forward.
As an engineer, I sometimes struggle with self-esteem issues and wonder if I've missed the AI development lifecycle and underestimated how fast things will change. However, after reading ‘The Godfather of A.I.’ Leaves Google and Warns of Danger Ahead, I am at peace. While many Twitter gurus, LinkedIn pandits, and some real AI experts may claim they saw it coming a long time ago and that AI is going to change the world, the godfather of AI also didn't see it coming. Looking at Paul Graham's essay to Chamath's podcast, if most people missed this trend, who am I to worry about it? I can focus on building things and stop worrying.
One quote that always rings in my mind is "Don't badmouth your competitors just because they're old and legacy companies with bad products. They solve real problems, so people use them even when their products are outdated." Will the AI race be like Slack vs. Teams? I haven't heard a single person praise Teams, yet they have 10x more daily users than Slack. True, Microsoft has failed many times, but they are experts in the enterprise market. No one gets fired for buying IBM; they have shifted to Microsoft.
What is the purpose behind all of these thoughts?
To be honest, I am not entirely sure. I have pondered this question for a long time but have yet to arrive at a concrete conclusion. While blockchain and web3 have been trending topics, I have dabbled in a few hobby projects but have not given them much serious consideration. However, with the recent advancements in AI and the way it has become a part of my daily operations, not just as a fancy tool but as a means of simplifying my life by learning things and solving problems for me, a new world has opened up. Perhaps it is because of this that I have started thinking more about the potential impact on the business world, which remains to be seen. At Docsumo, my job is to guide the team and solicit their feedback on how we can use AI to solve real-world problems, not just for demos. Finding real problems to solve is still human work, but AI can help us execute quick projects. However, if we want to get the most out of it, we need to collaborate with it, identify its strengths, work around its limitations, and strive to be 5–10 times better than we are currently. One thing is for certain: it can make us 10 times better developers, 5 times better product creators, and 2-3 times better at other things. Our job is to consume information, run experiments, stay informed, iterate based on data, plan for the future, work with what we have now, and build things that solve real problems, not just show demos on social media.
2. Podcasts/Essays
Here is a long article I recommend to anyone interested in startup.
This is an article I read at least once every six months (even before Sam Altman became super famous due to OpenAI). Helps me realign my thoughts.
3. Interesting links
Repos:
2048 chatgpt: Chatgpt plays 2048, the prompt is interesting in this project.
pdfGPT: We are closely monitoring the current development of large fundamental models. Large PDF files are one of the challenges for these models, but a lot of progress is being made in this space. Here is another good article on the same topic: Fixing Hallucination with Knowledge Bases.
Bark: I have two hobby projects that have been on my backlog for a long time. The first is data crawling and information consumption, and one aspect of this is text-to-speech. While Bark's quality is really good, it's also very slow. I'm looking forward to the day when text-to-speech can be run on a CPU with good-quality output.
Open-LLMs List: LLMs are all licensed for commercial use (e.g., Apache 2.0, MIT, OpenRAIL-M)
For more follow me on Github: bkrmdahal
Articles:
Rules of Thumb for Software Development Estimations: Estimating is difficult, but it is a skill set that everyone needs to learn. You get better with experience. All estimates are wrong, but you should be useful. Each post of Vadim is a gold mine.
Beware of AI pseudoscience and snake oil: I believe that something fundamental is indeed changing in AI, but being a bit more cautious will not hurt anyone. From the Article “I don’t think AI system vendors are lying. They are ‘true believers’ who also happen to stand to make a lot of money if they’re right. There is very little to motivate them towards being more critical of the work done in their field”
Google "We Have No Moat, And Neither Does OpenAI": A good timeline of how things are progressing in the AI race with open source breaking a lot of boundaries and how it can empower anyone to build the next big company with their own hands.
4. Quotes/ Books
A quick word about competitors: competitors are a startup ghost story. First-time founders think they are what kill 99% of startups. But 99% of startups die from suicide, not murder. Worry instead about all of your internal problems. If you fail, it will very likely be because you failed to make a great product and/or failed to make a great company.
- Sam Altman,
People tend to complain when things are not going well. The philosophy that always keeps me calm and in control is to never blame anyone, whether in a startup or in life. If Docsumo is successful, it's because the team was great and we executed well. If we are not doing well, then it's because we stink at decision-making and execution. It's not because someone left, people didn't do what they were hired for, or I have a small team for X, competitors, or the market. There are no other excuses except that my co-founder and I stink and need to improve. When this is the final conclusion, we are in total control and always steer the ship in the direction we want. There will never be a phase in life where we are more in control than owning the company with a war chest in the bank and a key in our password manager. :-D.
Our Slack channels have a great collection of memes.
That’s it for this edition. I hope you find it useful.
P.S. For now it will be a monthly edition. Setup in the USA and other tasks are taking a lot of time.
Best,
Bikram Dahal
P.S. If you learned something new today, please share “Notes from a CTO” with your friends and spread the love. ✌🏻
Wonderful writeup
Really nice article.