OpenAI CEO Sam Altman says AGI would have ‘whooshed by’ in 5 years with “surprisingly little” societal change than speculated

What you need to know

  • OpenAI CEO Sam Altman recently indicated that the company could hit the AGI benchmark within the next 5 years.
  • Interestingly, the executive claimed the feat would have “surprisingly little” impact on society.
  • This is amid AI’s exorbitant computing power and cooling water demands, prompting the AI firm to burn through large sums of money quickly.

As someone who has spent countless hours watching sci-fi movies and reading books about the rise of artificial intelligence, I must say, it feels like we are finally stepping into the realm of reality. The news that OpenAI could hit the AGI benchmark within the next five years is nothing short of mind-blowing.


Over the past few months, OpenAI has attracted significant attention, particularly following bankruptcy rumors, and forecasts suggesting potential losses amounting to $5 billion over the next year. Yet, the company behind ChatGPT managed to extricate itself from a tricky predicament by securing a round of investment, raking in a substantial $6.6 billion from investors such as Microsoft and NVIDIA. This infusion of funds elevated its market value to surpass $157 billion.

Sam Altman has passionately spoken about AGI (artificial general intelligence) and indicated that OpenAI is actively pursuing the audacious feat. In a recent Reddit AMA session conducted by the AI firm, the CEO stated that it’s possible to achieve the benchmark with current hardware.

Although the CEO didn’t explicitly detail the specific requirements, he earlier estimated that realizing his ambitious AI vision would involve investing around $7 trillion and a significant amount of time to construct approximately 36 semiconductor facilities and additional data centers.

It’s important to note that Artificial Intelligence (AI) currently demands a significant amount of energy and resources. Some experts are concerned that by 2025, there might not be enough electricity to sustain its rapid advancements, given AI’s potential role in driving one of the largest technological leaps humanity has ever seen.

In a nutshell, the electricity needs of tech giants like Microsoft and Google exceed the energy consumption of over a hundred nations, and this figure is expected to rise as technology progresses. Moreover, it’s been discovered that the technology’s cooling water demands are substantially higher than initially estimated, approximately four times greater.

AGI could be here sooner than we thought

Previously, Sam Altman estimated through a blog post that superintelligence could be within reach in a few thousand days. However, it appears that his estimated timeline for the development of Artificial General Intelligence (AGI) might now be more specifically set at 5 years, according to @tsarnick on X.

During a chat on the 20VC podcast hosted by Harry Stebbings, OpenAI CEO Sam Altman discussed future possibilities of artificial intelligence, expressing his views and forecasts.

In about five years, it seems that technological advancement will be happening at an astonishingly fast pace. Folks often remark that the moment of Artificial General Intelligence (AGI) arrived and moved on so quickly. The speed of progress is simply mind-blowing, and we’re constantly unearthing new information – not just about AI and research, but also across all areas of science as well.

Despite the widespread belief that advancements in AI and AGI will have a major impact on our society, Sam Altman suggests the opposite – that the effects might be smaller than people anticipate. To illustrate this point, he uses the example of computer capabilities: if you asked people five years ago whether computers would pass the Turing test (a test to determine if a machine can imitate human intelligence), they would probably say no. However, computers did indeed pass this test, but it didn’t result in significant societal changes – it went largely unnoticed.

Over a prolonged period, Sam Altman acknowledges that the AI advancement might trigger profound shifts in society, surpassing even our greatest predictions for scientific development. Notably, a previous researcher at OpenAI hinted that the company was close to achieving Artificial General Intelligence (AGI), but cautioned they weren’t adequately prepared or resourced to manage all its implications. This revelation follows another AI researcher’s claim of a 99.9% likelihood that AI could ultimately bring about human extinction, based on the p(doom) probability.

Read More

2024-11-05 12:39