As much as I want to be optimistic about AGI, I’m getting a little worried about it actually being developed

Over the last 3 years I was optimistic about AGI, but currently I’m getting concerned that it might not be developed in the 2030s (or even ever).<br><br>One of the first good arguments was the current computing power. <br>- Today, we have a $30 billion budget to train large models (but that might not be enough, especially with the competition in the AI market).<br>- In 2025, OpenAI promised to spend 90 billion on AGI development, but that’s based on strong revenue growth and current investors’ willingness to support the company. But there’s a lot of competition in this market and people who are investing in AGI might pull out.<br><br>The second argument was that we relied too heavily on large language models.<br>- While I’m impressed with ChatGPT 4 Turbo, the progress in its development was slowed down between ChatGPT 4 and ChatGPT 4 Turbo, and I’m not sure that we can expect this to change in the near future, especially with the growing competition. I’ve tried a few GPT-based models from other companies and they were on par with, or better than ChatGPT 4 Turbo in different areas. It seems like the competition is stifling the growth of these models.<br><br>The third argument was that people who are sceptical about AGI have recently come up with new strong arguments that no one could refute.<br>- For example, it has been argued that current models are far away from AGI and that general intelligence requires many more breakthroughs and discoveries to be made. But we’ve heard a lot about the progress in neuroscience and the recent discoveries that are promising, but it’s still in its infancy.<br><br>Overall, it’s been a pretty tough 6 months for the AGI community. I still hope for the best, but it’s getting tougher and tougher.