Suggested readings week 19/2023
The art of starting a business
To start a new company you need five things in place.
Mediocre teams do not build great companies
Putting idea, market, team, product and execution in the right order of relevance is key to success. Underestimating even one of these is the first reason for failure or low performances.
Sam Altman wrote a list of suggestions for startups: clearly stated, simple and effective in all businesses.
https://playbook.samaltman.com/
Artificial Intelligence for Dummies
Not really Artificial Intelligence, but Machine Learning.
https://www.parand.com/a-completely-non-technical-explanation-of-ai.html
Large Language Models are about to shrink
Large Language Models (LLM) have been a breaking revolution during the last couple of years. As an IEEE member said during an interview, it was as if Computer Graphics switched from Pong to Avatar in a matter of one week. The reaction from people across research, industry and society is understandably nervous.
The London tech boss says he is “incredibly nervous about the existential threat” posed by AI, even as he pursues it, and is “speaking with [other] founders about it daily”.
It's beneficial to the discussion to understand exactly how these computation models work and the extent of their usefulness. A collection of essays published by The Economist is able to give a correct overview in a simple language, and the seminal paper that came right after OpenAI reached extraordinary results with GPT-3 adds plenty of technical details.
Great to see that research on this topics is absolutely open and collaborative, thanks also to the visionary approach to business from the OpenAI founders. But there's a problem: the stock of high-quality language data will be exhausted soon, likely before 2026. What's next? Researchers and industry leaders are starting to envision a near future of totally different algorithms.
https://arxiv.org/abs/2005.14165
https://arxiv.org/abs/2211.04325
Closing roads
Closing a road can reduce traffic. And there’s a mathematical explanation for why.