Happy New Year 2024 and Year in Review 2023

Happy New Year Everyone!

I hope everyone has a great year in 2024!

What We Are Leaving Behind

I’m acknowledging but not going into details of the cruelty we are inflicting on ourselves in Ukrain, the Gaza strip, Myanmar, Niger, Burkina Faso, Azawad, Tunisia, Sudan, Columbia, Afganistan, Pakistan, Somalia, Democratic Republic of Congo, Nigeria, Iraq, Rwanda, Burundi, Mexico, Camaroon, Chad, Syria, Yemen, Saudi Arabia, United Arab Emirates, Ethiopia, Haiti, and pretty much everwhere.

Instead let’s skip over to the the weirdness that happened in the tech sector.

Generative AI Has Entered The Chat

Educated people often make the mistake of underestimating the disruption that can happen in an industry. In this case, we are talking about generative AI and its disruption of the tech sector.

Transformer models have been around for a while (since before 2017) but they have mostly been a curiousity rather than a serious practical tool. It held much promise and excitement but was considered to be a technology that was in its infancy. The only real production workloads they handled were around machine translation.

However, a sudden disruption happened in 2023 where:

• AI models become capable enough to give a somewhat convincing approximation of intelligence. This led people to conclude that generative AI is now ready for use in industry – a claim that is debateable.
• Generative AI models became accessible to the general public largely in the form of ChatGPT though others followed quickly.¹
• People started incorporating generative AI into their lives and businesses.²³⁴
• This triggered an influx of financial investment in generative AI.⁵⁶
• Which was followed closely by legislation that aimed to mitigate the perceived harm of AI.⁷

All this in the span of a year.

I’m not directly involved in any ML work, so the way I experienced it was:

• At the start of the year I had a couple of NLP projects that I had lined up to complete by the end of the year. The tasks were somewhat complex and involved fiddling with a lot of NLP.
• By the end of the year the tasks were reduced to producing a prompt for an LLM and then parsing its result.

In addition to all the assistive coding tasks that are pretty much production ready at this point, the transformation of a complex coding task into simple English question and answer is kind of mind blowing.

I’m waiting to see what the new year is going to bring in terms of automation. We are truly living in the future.