AI or Artificial Intelligence has been the tech buzzword since 2022 when ChatGPT 3.5 was released to the public. Before that, we have ML or machine learning where the power of hardware is fast enough and the explosion of data made neural network algorithms can be bigger than before. The result of Machine Learning is that we get pretty good pattern recognition. ChatGPT took the world by storm because, for chatbox software, its resemblance to human creation is uncanny.

Predictions arise that we are closer to the Singularity - the moment that our artificial creation has surpassed humans. Technologists glee that the fact we are closer to AGI - Artificial General Intelligence where software becomes self-aware and can do many tasks humans can do. And big tech companies like Microsoft, Google, Facebook and now Apple all have jumped on the bandwagon of AI. But are AI is the future or it is a bubble currently waiting to burst?

Current State

Let’s talk about the current state of AI, or more specifically generative intelligence, where algorithms create a rain of texts and pictures based on a prompt.

Microsoft


In the world of AI, ChatGPT is the go-to product. Microsoft invested tens of billions of dollars here.

Microsoft put $10 billion into OpenAI in 2022 on top of a $1 billion infusion in 2019. OpenAI has been making waves in the generative AI space since the introduction of ChatGPT 3.5 in November 2022. ChatGPT is the cause of the wave of money pouring into AI-related companies like Nvidia and Microsoft. What ChatGPT 3.5 caught everyone by surprise was how well it understands your questions and how well it answers them. Sure there were problems like hallucinations where the chatbox will keep rambling on incorrect info.

ChatGPT 4 comes into the scene later with even more capabilities but still retains the same problem like hallucinations. In the world of AI, Microsoft, and ChatGPT is the 800-pound gorilla. Most AI help tools are ChatGPT wrapper apps. What it means is that it just sends prompts to ChatGPT and prints out the output through an app.

Google


Google panicked with OpenAI features, which they have the right to be since OpenAI is built upon Google's idea. One of their 'me-too' product is Google Gemini.
When ChatGPT comes out, it sends shockwaves to Google. For the first time in a very long time, Google sensed an existential threat and went into panic mode. Google founders who are just about to walk into the sunset into retirement and philanthropy have been recalled to fight the good fight. Google is scared and companies like Google do not get scared.

Apple


The best marketing move by Apple is being late into the game, but redefine a well known phrase into your own. AI is Apple Intelligence in their world.

Apple is the technology company that will be at the forefront of technology but oftentimes, they will be the last one to join the party. However, when Apple decides to join a trend, it will always make an entrance and everyone will know that Apple has arrived.

AI is no different. Apple spent a better part of an hour in WWDC 2024 explaining new AI features on the next version of operating systems that will run on all Apple devices. Apple took the next step in marketing, or better yet, the genius step in rebranding AI as Apple Intelligence instead of the ubiquitous Artificial Intelligence. It’s AI with Apple’s fingerprint all over.

As of press time, Apple’s AI feature hasn’t made it in the wild, but when Apple jumped on the bandwagon, it meant that AI was not a fad, but a feature that will stay with us for the next few years.

Nvidia


Nvidia's best pickaxe to solve the AI problem. It's very expensive and sales has been restricted due to government strategic concerns.

While Nvidia itself is not known for making AI models, they are the ones who are making the hardware to make AI models possible. In a gold rush, the one who stands to gain is the one who is selling pickaxes to the gold miners. And Nvidia has the best pickaxes in the industry.

Nvidia might not make splashing on the front page by introducing a new AI model that can make cat videos, or chat to you like a real human, but their cabinet-sized AI accelerator has been selling like hotcakes to big tech companies like Facebook and Microsoft. For some, Nvidia is what Cisco was during the dot com boom years. It’s both an aspiration and a cautionary tale.

Problems

AI feature seems to dazzle a lot of users and investors, but they are not without problems.

Energy consumption is a major problem. Google has pledged to be fully 100% green energy by 2030, but power-hungry AI modeling data centers aim to kill that goal. Google emissions climbed 50% in 5 years because of AI demand. The same problem is occurring for Microsoft too.

For comparison, the human brain uses around 20W of energy. Much lower most of the time. Sure, humans cannot calculate the product of two 12-digit numbers quickly, or draw pictures so fast that it looks like a movie, but the human brain is very efficient in doing tasks like picture recognition or writing a sentence.

The other problem is despite its widespread use, there’s no silver bullet solution involving AI … yet. Unlike putting a camera on a phone, or having a spreadsheet on a computer, the AI use case current is, as Apple has shown to us, a feature, not a product.


Instead of an AI chatbot, some companies experiment with connecting remote workers from low-income countries to be a virtual cashier over high-speed internet.

There’s also the competition. With super-speed internet, some companies experimented with bio-robots: low-wage workers that work in faraway countries and do the job below the minimum wage.

Another goal of AI researchers is to achieve AGI - Artificial General Intelligence. Humans might not be able to match the performance of AI models in certain tasks, but in general, they can pretty much outsmart a single-task AI model. For example, can you ask ChatGPT to change a flat tire for you? Of course not.

And there’s the problem of hallucinating. AI model problem is that: they are models. They would analyze your prompt and respond to it, but most likely, without guardrails, they would not know whether the thing that they are doing is morally right or wrong. Or the information that is presented is factually correct or not.

Conclusion

Now the real question. Is AI a fad? Not really. AI features like helping you to draft a letter or even write code are here to stay and I suspect it will get better through time. According to a Stanford study, investments in AI-centric companies are still growing. It also could mean that we haven’t found anything that is new and more engaging. Google Trends report shows that AI as a search term is still popular compared to past hypes like crypto or NFTs.

Plug

Support this free website by visiting my Amazon affiliate links. Any purchase you make will give me a cut without any extra cost to you

BasePro
iPhonesiPhone 16 / iPhone 16 Plus - (Amazon)iPhone 16 Pro / iPhone 16 Pro Max - (Amazon)
WatchApple Watch SE (Amazon) / Apple Watch Series 10Apple Watch Ultra 2 (Amazon)
AirPodsAirPods 4 (Amazon)AirPods Pro 2 (Amazon) / AirPods Max (Amazon)
iPadiPad 10 (Amazon) / iPad Mini (Amazon)iPad Air M2 (Amazon) / iPad Pro M4 (Amazon)
LaptopsMacBook Air M3 (Amazon)MacBook Pro M3 (Amazon) / MacBook Pro M3 Pro/Max (Amazon)
DesktopMac Mini (Amazon) / iMac (Amazon)Mac Studio / Mac Pro
DisplaysStudio Display (Amazon)Pro Display XDR (Amazon)

Other Ecosystem Items