Transcription
I believe that someday we will make something that qualifies as an AGI. By whatever fuzzy definition you want, the world will have a two-week freakout, and then people will go on with their lives. Sam Altman just said the world will only have a two-week freakout when we get to AGI. That's quite a statement to make. One thing I say a lot is no one knows what happens next. Sam Altman just said the words no one expected.
OpenAI has been on the wrong side of history when it comes to open-source AI. This is a company that built its brand on pushing AI innovation forward, but now, even its CEO admits they might have made a mistake. So what exactly is happening inside OpenAI, and why is Altman reconsidering the company's approach? OpenAI's dominance is being challenged, and the cracks are starting to show. From rising competitors like DeepSeek to internal debates about open-source AI, something major is shifting in the industry. If OpenAI doesn't adapt, it could risk losing its position at the top. Let's break down exactly what Altman admitted, what's changed, and how this could reshape the future of AI.
During a recent Reddit AMA, Altman said, "I personally think we need to figure out a different open-source strategy." He added that not everyone at OpenAI agrees with him and it's not their highest priority right now, but the fact that he's acknowledging this at all is significant. For years, OpenAI positioned itself as the leader in AI research, but it's never been a truly open-source company. Unlike Meta or Google DeepMind, which have released large-scale open models, OpenAI has kept its most advanced technology proprietary. That's why Altman's recent statement is such a big deal; he openly admitted that the company's stance on open-source AI was a mistake.
Why now? The pressure is mounting. OpenAI is facing criticism from researchers, the open-source community, and even its own users who believe that AI should be more accessible. But more importantly, there's a new wave of competition that's forcing OpenAI to rethink its approach. OpenAI struggled to maintain its lead for years. OpenAI had a comfortable lead in AI innovation, but that's no longer the case. Altman admitted that Chinese AI company DeepSeek has significantly narrowed the gap. DeepSeek recently launched its own reasoning model, R1, which directly competes with OpenAI's AI reasoning capabilities. Unlike OpenAI's models, which hide their thought process, DeepSeek's model is completely transparent, showing every step of its reasoning. That's a big deal because it allows researchers and developers to understand exactly how the AI reaches its conclusions—something OpenAI has intentionally kept hidden. OpenAI has even accused DeepSeek of stealing its intellectual property to catch up. While no concrete evidence has been made public, the allegations highlight just how serious this competition has become. And it's not just DeepSeek; companies like MRR AI, Anthropic, and Meta are all making rapid progress in AI research. Open-source models like Meta's Llama 3 are becoming powerful enough to rival OpenAI's closed-source models, raising the question: Is OpenAI's secretive approach still sustainable?
What are you excited about in 2025? What's to come? AGI? Yeah, uh, excited for that. Uh, what am I excited for? OpenAI's shifting open-source strategy. Historically, OpenAI has been very selective about what it releases. GPT-2 was open-source, but GPT-3, GPT-4, and GPT-4 Turbo were not. The company argued that keeping models closed was necessary for safety and competitive reasons. Now that mindset is starting to change. OpenAI's Chief Product Officer, Kevin W., said in the same AMA that the company is considering open-sourcing older models that are no longer state-of-the-art. This would be a major shift, but is it enough? If OpenAI only open-sources outdated models, that still doesn't address the demand for more transparency in its most advanced AI systems. And with competitors like DeepSeek making their models fully open, OpenAI might not have a choice but to adapt.
But there's another reason OpenAI has been hesitant about open source: security risks. Altman has repeatedly warned about the dangers of releasing powerful AI models without proper safeguards. If OpenAI were to make its most advanced models public, it could open the door to misuse, including the development of deepfakes, misinformation campaigns, or even autonomous hacking tools. So the big question is: Can OpenAI find a balance between openness and security?
The ChatGPT pricing controversy. Beyond the transparency debate, OpenAI is also facing pushback on its pricing strategy, especially with ChatGPT. Right now, ChatGPT offers a free plan, a $20 per month plus plan, and a $200 per month Pro Plan. But earlier this year, Altman admitted that OpenAI was losing money on its Pro Plan. This immediately raised concerns that prices could increase in the future. When asked about this in the AMA, Altman said he wants to make ChatGPT cheaper over time, not more expensive, but how realistic is that? Running large AI models is extremely costly. OpenAI is investing billions into infrastructure, including its massive Stargate data center project. With the cost of computing power rising, it's unclear how OpenAI can continue improving ChatGPT while also lowering prices. Some speculate that OpenAI might introduce tiered pricing or additional monetization strategies to offset costs, but for now, the company is staying quiet about any major changes.
The real reason OpenAI is pushing for more compute power. OpenAI's ambitions aren't just about better AI; they're about raw computing power. The company is currently building Stargate, a massive AI data center project designed to house the next generation of AI models. But why does OpenAI need this level of infrastructure? Kevin W., OpenAI's Chief Product Officer, explained it simply: more compute power equals better AI models. But it's not just about improving ChatGPT or doing AI; the industry is in an arms race, and OpenAI is competing with Google DeepMind, Meta, and Anthropic to develop the most advanced AI systems ever created. This competition isn't cheap. Training cutting-edge AI models costs billions of dollars in hardware, electricity, and data storage. Microsoft has already invested over $13 billion in OpenAI, with plans for even more funding to keep up with demand. But the real concern isn't just the cost; it's what this level of compute power enables. AI researchers have long warned that as models become more powerful, they can start improving themselves in ways we don't fully understand. This concept, known as recursive self-improvement, could lead to AI systems rapidly accelerating their own intelligence. If OpenAI achieves this before proper safety measures are in place, it could push the world closer to artificial general intelligence (AGI)—a scenario where AI reaches human-like or even superhuman levels of reasoning. AI experts have warned about loss of oversight, where AI systems develop strategies, behaviors, or decision-making processes that humans don't understand. And if OpenAI's next-generation models can already perform advanced reasoning tasks, what happens when they begin optimizing their own architecture?
These concerns aren't just theoretical. Governments and regulatory bodies have started paying close attention. In March 2024, the US and UK formed a joint AI safety task force to study the risks of uncontrolled AI development. Meanwhile, China has imposed strict regulations on AI model transparency, requiring companies to disclose training data and methodology. OpenAI insists that safety is its priority, but the bigger it gets, the harder it becomes to slow things down. Could OpenAI be developing AI too powerful for even its own engineers to control?
OpenAI's secretive role in nuclear research. While OpenAI publicly promotes responsible AI development, its latest partnership with the US government raises serious ethical questions. In early 2025, OpenAI announced a collaboration with the US National Laboratories to apply its AI models to nuclear defense research. This immediately set off alarms. AI and nuclear technology are a dangerous combination, and users in the OpenAI AMA wasted no time asking the obvious question: Could OpenAI's AI be used to develop nuclear weapons? Kevin W. responded by saying he trusted the government and that the scientists involved are AI experts who understand the risks, but for many, that wasn't reassuring. AI and military applications aren't new; governments have used machine learning for everything from missile guidance systems to cybersecurity threat detection. But OpenAI models are far more advanced than traditional military AI, and their potential impact is uncertain at best, catastrophic at worst.
What's next for OpenAI? Despite ongoing speculation, Sam Altman has confirmed there's no set timeline for GPT-5. However, with Google DeepMind, Anthropic, and MRR AI advancing rapidly, many expect GPT-5 could be announced before the end of 2025. OpenAI will need to deliver something groundbreaking to maintain its dominance. Meanwhile, OpenAI's Q* reasoning model is already here, aiming to improve complex logical tasks. This puts it in direct competition with DeepSeek's R1, which openly displays its thought process—something OpenAI has historically avoided. On the image generation front, OpenAI is working on a successor to DALL-E 3, promising it will be worth the wait. However, with Midjourney 6 and Stable Diffusion XL leading the market, OpenAI has work to do to catch up.
Beyond new models, OpenAI's rapid expansion raises bigger questions. Is the company spreading itself too thin? With growing competition, regulatory scrutiny, and internal debates, OpenAI faces mounting pressure. Altman has admitted OpenAI has been on the wrong side of history regarding open-source AI, but will this actually lead to transparency? So far, the company has only hinted at open-sourcing older models, leaving its most advanced systems closed. The biggest challenge may not be competition but keeping up with their own creations. As AI evolves at an unprecedented rate, OpenAI must decide whether to embrace openness or risk losing trust. Where OpenAI goes from here remains uncertain, but one thing is clear: This is just the beginning of a much bigger AI transformation. If you've made it this far, let us know what you think in the comment section below. For more interesting topics, make sure you watch the recommended video that you see on the screen right now. Thanks for watching.