2024 has been a frenetic year for AI with dozens of new models launched with increasingly sophisticated capabilities, but things might get a bit more challenging from here on.
Google CEO Sundar Pichai has said that progress in AI is going to get harder in 2025. “(For) the current generation of LLM models, a few companies have converged at the top, but I think they’re all working on (their) next versions, too,” he said in an interview. “I think the progress is going to get harder when I look at 2025. The low hanging fruit is gone. You know, the curve, the hill is steeper. I think the elite teams will stand out in 2025. So I think it’s an exciting year from that perspective,” he added.
Pichai’s comments were in contrast to what OpenAI’s Sam Altman seems to believe. Altman has said that there’s “no wall” in AI, implying that AI progress hasn’t hit a wall, and the current rate of progress could go on for a while.
Pichai was asked to comment on Altman’s views by the interviewer. “I’m very confident there will be a lot of progress in ’25,” he replied. “I think the models are definitely going to get better at at reasoning and completing a sequence of actions more reliably in a more agentic way. So I don’t fully subscribe to the wall notion. When you start out quickly scaling up, you can throw more compute and you can make a lot of progress. But you’re definitely going to need deeper breakthroughs as we go to the next stage,” he added.
“So you can perceive it as there’s a wall or you perceive it as, you know, there’s some small barriers,” Pichai laughed.
There has been some chatter about AI progress slowing down. People have noticed that the improvements in AI models is becoming incremental, and a step jump in capability — as was seen from GPT-3 to GPT-4 — hasn’t happened in a while. But most AI leaders have maintained that they believe that AI progress will continue unabated — NVIDIA CEO Jensen Huang has said that AI is progressing much faster than Moore’s law, and Anthropic CEO Dario Amodei has said that AI could lead to a century of scientific progress in the next 5-10 years. But Google CEO Sundar Pichai has struck a more cautionary note, saying that he believe that while there will still be substantial progress made in AI in the coming months, its rate might be slightly slower than what we’ve gotten used to in the recent past.