Grok Code Fast 1 has just been released and its usage is off the charts. It is offering agentic coding capabilities at a fraction of the price that other competitors are offering. There are many experts who claim that all these AI companies are actually running at a loss, and they are offering their products at throw-away prices or for free because they want people to learn it, use it, and consequently, become hooked to it by making it part of their workflow.
Though I do not believe in everything what is being said, I do acknowledge that AI companies are offering their products at heavily discounted prices for nuanced reasons that partially stem from the fact that to develop AI, more and more use-case scenarios are required. Teaching AI physics and math is great, but what is equally important is to have people use it. Without a diverse set of people using it, it is impossible for AI to develop in a general manner where it can claim itself to be the jack of all trades.
Also, it would truly suck if all the power of AI is reserved for only the rich and the privileged. It is a blessing in disguise that AI companies are forced to offer their products to the public, not that they are generating huge profits out of it, but in the hope of generating humongous future profits, these companies seem to believe that their best bet is on the people themselves.
Many AI companies like OpenAI, xAI, and Meta are on the verge of operationalizing their respective 1 Gigawatt+ energy consumption data centers. To set up these data centers alone, hundreds of billions of dollars are being invested. Ultimately, without even realizing it, the AI industry is slowly morphing itself into the Energy Industry and might even hijack the global energy industry in the next few decades. Famous computer scientist, Andrew NG, adjunct professor at Stanford University and founder of Coursera, believes that ‘AI is the new electricity!’
Such a tectonic shift will have global implications and what the future holds, nobody knows. But it surely does not look bleak. AI doomers may continue to point out that soon AI will take over everything, but that is definitely not going to be the case. There is a catchphrase that is roaming in the technology corridors these days that intelligence has been solved, and it is mostly a matter of energy. The more energy you give to these AI systems, the better they perform. I do not subscribe to such overly simplistic notions, but I do agree that we are living in an age of innovation and disruptive changes. A very exciting time to live in.
So, let the cost of computing and cost of using AI keep sinking. That is the best scenario for everyone. Moore’s law may be dead, but thankfully, the Huang’s law and Jevons paradox seem to be holding strong for now.