Catch Up with AI

Cover Photo by charlesdeluvio on Unsplash

It has been over 10 months since I last posted, so now it's time to catch up. I used to write about technical concepts from a fundamental point of view so that people with non-technical backgrounds could understand them. This led to a series of 11 articles on topics like full stack development, machine learning and artificial intelligence, cloud computing, and some other fundamental topics like blockchain and version control. They can still be accessed on my Medium profile.

Now, I plan on starting to write more about my experiences that I have had over the past few years being a student and professional in the tech industry and express my personal views on ongoing events in the industry. Let me start with a little background about myself. I am currently a student enrolled in an undergraduate degree in computer science, and I have been working in tech as an independent contractor for close to 5 years now. I am an AWS certified Solution Architect, and I specialize in developing data-intensive systems along with a few other fields. Enough about me, now let’s dive into what’s going on lately.

AI Hype and History

AI (or more specifically GenAI) is by far one of the hottest topics in Silicon Valley. The funding for startups involving the use of AI has been astronomically high. According to Scale Capital, while the global venture funding for startups saw an increase of just 3% from 2023 to 2024, the funding for AI startups saw an increase of 80%.

In my case, the implications of GenAI are more directly seen in two popular scenarios: In college, by the students (including me) to complete assingments. In the industry, by startups trying to add GenAI to their existing products.

Now AI itself is as old as computers, and ML advancements have been taking place for a very long time. AI-based virtual assistants like Amazon Alexa, or Apple’s Siri have been around for over a decade now. Email spam filters, which most of us use, have also been using ML since the early 2000s. So what’s causing this hype now? I think two major factors are: Accessibility of GenAI and Scalability of GenAI.

Before 2017, GenAI was not scalable since it took a lot of computation power and was not parallelized but after researchers at Google introduced the transformer architecture in 2017, the GenAI became much faster to train and hence much more scalable across fields. Building the GenAI model still takes a lot of computation, but now it does not take as much time. This led to further enhancements like the introduction of the BERT and GPT series in the following years.

But still, it was not very accessible to the general masses until the ChatGPT model went viral in 2022. Since for most of the masses, the concept of a fundamentally intelligent model was very unexpected, this started the hype around what people nowadays call AI. ChatGPT was and is an excellent model, but people tend to forget that it is essentially an autocomplete tool with the capabilities of a search engine.

Current Economic Stance

Circling back to the industry implications, most startups started adding GenAI-based features in their products to catch up with the public’s demand. Even though many of these features are very useful, I think the majority are unnecessary additions. Now, I could be completely wrong about this since at one time, e-commerce was also considered unnecessary by many, and today it is one of the leading industries. But even with e-commerce (ecom), the market hyped it too much, and countless people created their personal stores hoping to generate high revenues who ended up spending high advertising bills with negligible online sales. We will table the discussion of the social impacts of ecom for some other post. I on many occasions forecast the hype of AI to have similar trends to that of Blockchain (Cryptos or NFTs) in 2021. According to Pitchbook, Over $30B of venture funding was deployed in crypto startups in 2021. That bubble burst in Q2 of 2022 with more than an 80% dip in funding compared to Q1 of the same year. Again, this does not mean that Web3 or Blockchain is a bad technology, or that it is a scam. All I am trying to say is, that it was overhyped and hence many many engineers hired for those startups on astronomically high packages during the hype were laid off in the following years. I think a similar trend will be seen for GenAI startups. Right now, AI startups — hiring engineers — that are trying to fit in AI where it is not really needed will soon realize how the liquidity is not high enough for them to exit.

College Life

I have been wanting to write on this topic for a very long. This topic even deserves its own article. The college experience is so complex. The networking and social aspects, the classes and academic aspects, the exposure and opportunities, the obstacles and challenges. But focusing this post more towards AI, let’s talk about the academic aspect.

I consider college to be a great place since you not only learn from your classes but also develop very strong and helpful networks among your peers. But lately, the learning part of the classes has been negligible for most of the students around me in computer science. The majority can get solutions for their assignments with minimal effort using AI, and most of them are essentially not understanding anything. I am a strong advocate of the ideology that “if you can find answers in your classes, you can find answers in your jobs.” But that is only valid as long as one can understand what is happening. If you find the answers to your assignments without giving any thought to what the problem asks for and how the answer answers it, it probably in my opinion is not going to benefit anyone. And when faced with a challenge that requires an actual understanding of the tech concepts, merely finding answers might not be the best skill. It has to be accompanied by the skill of knowing how to use those answers to solve problems.

So, when students who barely own any fundamental technical knowledge go out into the industry, they will struggle as they are neither employable nor coachable. This is what I have been noticing as well — there are plenty of recruiters saying “there is no talent available to hire” and plenty of graduates saying “there are no jobs available.” I think just developing some fundamental technical skills will go a long way in the future since the other competing professionals will lack that.

Final Notes

I do not mean to look down upon the AI industry, on the contrary, I see great potential in it. I think developing relevant skills in this sector can go a long way, but if this hype continues the bubble will keep getting bigger and bigger. If and when the bubble bursts, the industry might find itself in a strong recession. I do not reflect anyone else’s opinion but mine (with some conviction though).

I hope you enjoyed the read. Please share your opinion on this even if it does not match mine.

Feel free to connect with me on Linkedin.

This post was originally published on Medium.