Artificial General Intelligence (AGI) and the Singularity have been hot topics in AI for years, sparking both fear and excitement. However, experts like Sam Altman predict AGI will arrive in 2025, while Elon Musk predicts it will happen in 2026. But are these predictions based on reality or just hype?
In reality, there won’t be any AGI by 2025. Instead, large language models (LLMs) will find their “killer app.” To understand why LLMs have significant shortcomings and aren’t suitable for achieving AGI, let’s look at the technology.
Sentence completion is not intelligence or AGI. Generative AI, like OpenAI’s GPT models, can hold human-like conversations but only by spotting and repeating patterns. These systems don’t truly “consider” other options and lack self-awareness. For example, when given the input “Life is like a box of…”, the prediction might be “chocolates,” but it doesn’t consider other options like “box of surprises.”
OpenAI’s O1, released in 2024, does not directly answer questions but creates a plan to determine the best way to answer them. This chained output is impressive, but it’s still a fixed framework and not dynamic. While this approach shows promise, it highlights the fundamental technological challenges we face in achieving AGI.
The barriers to getting to AGI include:
* AI struggles with context and often misses important details humans pick up naturally.
* Current AI builds outputs on previous ones, so mistakes can snowball.
* Many researchers believe that much of these issues might be fixable, but not this year.
In 2025, we’ll see more narrow AI solutions integrated into chains similar to OpenAI’s “o1” approach. These systems will excel at specific tasks and enhance productivity and surpass human performance in many areas. While exciting, these advancements won’t constitute AGI. The focus should be on the real risks and opportunities of AI rather than getting sidetracked by AGI debates.
Sam Altman’s claim of AGI might be partly marketing-driven, grabbing attention with bold claims. Instead of focusing on AGI, let’s look at the biggest applications of large language models in 2025. Stay tuned for my next prediction, which will explore this further.
Source: https://www.forbes.com/sites/lutzfinger/2025/01/03/no-agi-but-a-killer-app2025-ai-prediction-110