GPT-5 Coming In December

0:00Today on the AI Daily Brief, AI is officially a national security issue. and before that in the headlines, GBT5 appears to be coming by December. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes. Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. Although today we are doing an open AI-centric headlines and there may be more than five minutes as there is a lot of news that we have to catch up on from the last couple of days, starting with the biggest news, reports that GPT-5, which has been codenamed Orion internally, is coming this year by December.

0:40The Verge reports that OpenAI plans to launch Orion by December, although other sources say that engineers inside Microsoft are preparing to host Orion on Azure as early as November. This would, of course, be the two-year anniversary of ChatGPT, so it makes it a fitting time. A note on naming, they say that while Orion is seen inside OpenAI as the successor to GPT-4, it's unclear if it will actually call it GPT-5 externally. Another interesting note from the Verge reporting is that unlike most of OpenAI's models, including the last two that have been released, GPT-4.0 and O1, Orion is not going to initially be released through ChatGPT.

1:15Instead, it is going to be released through the API first to a select group of company partners. Now, one of the important things here is that O1 was not a bigger model than GPT-4.0. It just took a different approach, which has made it better at certain reasoning tasks. Orion, which we might think about as GPT-5, is reportedly a much more advanced model. One OpenAI executive teased it as potentially 100 times more powerful than GPT-4. Back in September, apparently OpenAI researchers threw a happy hour to celebrate finishing training Orion. And around the same time, OpenAI CEO Sam Altman said, I love being home in the Midwest.

1:51The night sky is so beautiful. Excited for the winter constellations to rise soon. They are so great. The winter constellation that is most visible in the night sky from around November to February is, of course, Orion. Now, obviously, there are big things riding on this model. OpenAI has, on the one hand, recently closed a huge round of funding at a massive valuation with big plans to convert the nonprofit into a for-profit structure. But at the same time, it's coming up on a year of near-constant controversy. Firings, rehirings, executive departures. Now, obviously, investors have a lot of faith in OpenAI's ability to deliver given its$157 billion valuation.

2:27But GPT-5 or Orion or whatever it ends up being called has some big shoes to fill. Perhaps the other biggest news from OpenAI this week is that the company is winding down its AGI readiness team. The information reports that Miles Brundage, an OpenAI policy leader focused on safety, announced on Twitter that he would be leaving the company so that he can, quote, have more freedom to publish research and be more independent. The key points that he shares are that, one, he thinks he'll have more impact as a policy reacher and advocate in the nonprofit sector. Two, he argues that, quote, working at OpenAI is an incredibly high-impact opportunity now more than ever.

3:01OpenAI needs employees who care deeply about the mission and who are committed to sustaining a culture of rigorous decision-making about development and deployment. The precedents we set with each decision and launch matter a lot. Three, he left a reminder that his colleagues need to speak up when they disagree with decisions being made by executives. He writes, I think people almost always assume that it's harder and more costly to raise concerns or ask questions than it actually is. OpenAI has a lot of difficult decisions ahead and won't make the right decisions if we succumb to groupthink. Now, Brundage has been with OpenAI for about six years.

3:29In a substack, he elaborated on his thoughts around AGI, writing, In short, neither OpenAI nor any frontier lab is ready, and the world is not ready. I think that improving frontier AI safety and security is quite urgent given the number of companies dozens that will soon, next few years at most, have systems capable of posing catastrophic risks. There are also difficulties around credible commitments to and verification of safety levels, which further incentivized corner-cutting. But lest you think this was all doom and gloom, he also said, I think it's likely that in the coming years, not decades, AI could enable sufficient economic growth that an early retirement at a high standard of living is easily achievable.

4:02Before that, there will likely be a period in which it is easier to automate tasks that can be done remotely. He gets philosophical, writing, In the near term, I worry a lot about AI disrupting opportunities for people who desperately want work, but I think it's simultaneously true that humanity should eventually remove the obligation to work for a living and that doing so is one of the strongest arguments for building AI and AGI in the first place. The now-defunct AGI readiness team will be distributed to other divisions. A spokesperson for OpenAI said, Brundage's plan to go all in on independent research on AI policy gives him the opportunity to have an impact on a wider scale, and we are excited to learn from his work and follow its impact.

4:33We're confident that in his new role, Miles will continue to raise the bar for the quality of policymaking in industry and government. Moving on to the next OpenAI story in this longer-than-normal headlines, Microsoft and OpenAI are partnering to give news outlets millions of dollars to use AI tools. The grants are split evenly between cash and software credits and are focused on large regional news outlets. The grants will fund a two-year fellow at each outlet to work with them in developing and implementing AI tools. And the initiative is being conducted in collaboration with the Lenfest Institute for Journalism, which works to promote local media.

5:04Tom Rubin, the head of IP and content at OpenAI, said in a press release, while nothing will replace the central role of reporters, we believe that AI technology can help in the research, investigation, distribution, and monetization of important journalism. A small one, but great for the folks in Europe, OpenAI has now rolled out its advanced voice mode to EU users. Announcement pretty surreptitiously, when someone tweeted, Any update for us in Europe? They said, well, yes, all plus users in the EU, Switzerland, Iceland, Norway, and Liechtenstein now have access to advanced voice. The company did not say anything else about how their calculus around EU regulations had changed, but as I said, great for European AI users who have access to those features now.

5:40Finally, if we have one story about OpenAI executives leaving, we have another about new executives joining. During the Biden years, Chatterjee was credited with helping coordinate the implementation of the CHIPS Act, which authorized roughly$280 billion in spending. OpenAI said that Chatterjee will be tasked with research including, quote, the global economic impacts of building AI infrastructure, insights on longer-term labor market trends, and how to help the current and future workforce harness the benefit of this technology. The other big hire announced this week was Scott Schools as the company's first Chief Compliance Officer.

6:08Schools held the same role at Uber from 2018 and was previously the highest-ranking lawyer at the Justice Department. OpenAI finds themselves in a similar position that Uber was in six years ago, needing to keep up with changing legal frameworks in multiple jurisdictions. Che Chang, OpenAI's general counsel, said, Scott's deep expertise will further strengthen our team's ability to deliver beneficial AI technology while continuing to operate with the highest integrity standards and adapting to rapidly evolving regulatory environments. That is going to do it for our very OpenAI-centric headlines today.

6:36Next up, the main episode. Today's episode is brought to you by Venice. Venice is a private, uncensored generative AI app. It accesses open source models to enable text, image, and code generation without the fear of being spied on or having your data exploited. Discuss anything with Venice without concern about it being monitored, sold, or given to advertisers and governments. Venice is different because your conversations and creations are kept securely within the browser, never stored or accessible by Venice. Unlike other AI apps, Venice won't tell you what's okay to say or not. Venice won't patronize you.

7:07It simply provides direct access to machine intelligence. No topics are off limits, no ideas are taboo. With Venice, you're in control of the AI as you should be. Pro subscriptions are available for$49 a year or$8 per month. AI Daily Brief listeners receive a 20 % discount on Venice Pro. Visit venice.ai slash NLW and enter the discount code NLWDailyBrief. That's NLWDailyBrief, all one word. Today's episode is brought to you by Superintelligent. Every single business workflow and function is being remade and reimagined with artificial intelligence. There is a huge challenge, however, of going from the potential of AI to actually capturing that value.

7:47And that gap is what Superintelligent is dedicated to filling. Superintelligent accelerates AI adoption and engagement to help teams actually use AI to increase productivity and drive business value. An interactive AI use case registry gives your company full visibility into how people are using artificial intelligence right now. Pair that with capabilities building content in the form of tutorials, learning paths, and a use case library, and Superintelligent helps people inside your company show how they're getting value out of AI while providing resources for people to put that inspiration into action.

8:19The next three teams that sign up with 100 or more seats are going to get free embedded consulting. That's a process by which our super intelligent team sits with your organization, figures out the specific use cases that matter most to you, and helps actually ensure support for adoption of those use cases to drive real value. Go to bsuper.ai to learn more about this AI enablement network. And now back to the show. Welcome back to the AI Daily Brief. Today we are checking in on AI and the government, and frankly, we haven't had a lot to talk about in this domain recently. Partially that is, of course, because of the election cycle, kind of dominating everything in politics.

8:56But this week, the White House has made AI advancement a national security priority. It defined AI as an era-defining technology. The landmark National Security Memorandum directs government agencies to step up experiments with and deployments of AI. During a press briefing on Thursday, National Security Advisor Jake Sullivan said, This is our nation's first ever strategy for harnessing the power and managing the risks of AI to advance our national security. He commented that the speed of development in AI is quote-unquote breathtaking, and that it has the potential to change all manner of fields from nuclear physics to rocketry.

9:27According to unnamed White House staff, one of the key goals is to set clear guidelines on adoption. They said, We must out-compete our adversaries. With a lack of policy clarity and legal clarity of what can be done and can't be done, we are likely to see less experimentation. With that in mind, the memorandum has some clear guardrails. It says that AI should not be used to track Americans' free speech or get around existing controls on nuclear weapons. It drew a line in the sand, in fact, that AI must never be placed in control of the launch of nuclear weapons. More generally, it says that the technology must not be used in ways that, quote, do not align with democratic values.

9:57However, it is very clear that the overall tone is one of acceleration. During a press briefing, a senior White House official said the U.S. currently has a very strong hand in AI. The nation is currently leading on designing advanced hardware and building advanced AI systems while catching up on domestic chipmaking. At the same time, the White House noted that innovation is largely driven by the private sector. It needs to be followed by the government, particularly the national security agencies. They said, a failure to do this, a failure to take advantage of this leadership and adopt this technology, we worry could put us at a risk of strategic surprise by our rivals such as China.

10:27The unclassified parts of the memo featured a number of extremely important policy changes. It is now official U.S. policy to, quote, lead the world's development of safe, secure, and trustworthy AI. To that end, the government is directed to promote and secure foundational AI capabilities. The White House said that U.S. leadership cannot be taken for granted and must be quote-unquote proactively strengthened. The U.S. government is now directed to harness powerful AI to achieve national security objectives. The memorandum noted that employing these systems effectively would quote, require significant technical, organizational, and policy changes.

10:56Safety is a heavy emphasis with the newly established AI Safety Institute directed to provide a host of risk and safety assessments over the next nine months. Another key policy change is that advancing the ability for quote, non-citizens highly skilled in AI and related fields to enter and work in the United States constitutes a national security priority. Defense and Homeland Security Departments are directed to assist in rapidly bringing these workers into the United States. While this section seems to be mostly about the private sector workforce, it also implies more rapid vetting and authorization for non-citizens to work on sensitive military projects.

11:25Defense, Energy, and Commerce Departments are now directed to make public investments and encourage private investment in strategic domestic and foreign AI technologies. This section has a large focus on computational infrastructure. Relevant agencies are now directed to streamline the permitting, approval, and incentives for AI infrastructure, including energy. The memorandum specifically includes nuclear energy as an approved clean energy source and mentions power transmission lines as a focus. The government has also been directed to guard against the risks of foreign actors obtaining U.S. intellectual property through technology transfers and data localization requirements.

11:55Collection of intelligence on corporate sabotage and infiltration of the AI sector is now a top-tier priority for intelligence agencies with a view to securing innovations. Now, there are dozens of other directions, but the main point is that leadership in AI and rapid acceleration of AI adoption is now a national security priority. Throughout the memorandum, there was an overtone that the U.S. needs to step up and demonstrate the way to adopt AI without jeopardizing democratic norms. The memorandum stated, If misused, AI could threaten United States national security, bolster authoritarianism, undermine democratic institutions and processes, facilitate human rights abuses, and weaken the rules-based international order.

12:31Now, OpenAI, for their part, have already weighed in on the memorandum, providing their response in a blog post. They wrote, We believe a democratic vision for AI is essential to unlocking its full potential and ensuring its benefits are broadly shared. AI is a transformational technology that can be used to strengthen democratic values or undermine them. That's why we believe democracies should continue to take the lead in AI development, guided by values like freedom, fairness, and respect for human rights. And it's why we think countries that share these values should understand how proper safeguards AI can help protect people, deter adversaries, and even prevent future conflict.

13:02They listed a handful of ways the company is already engaging with government projects from cybersecurity to bioscience research. However, they noted we need clear guardrails and policies around how AI can be used. That's why we're taking a careful measured approach to our national security partnerships. They continued saying that the new White House framework, quote, opens up the potential to support more national security work in the U.S. and allied countries in a way that stays true to our mission. Ultimately, if last year's AI executive order was about the government figuring out how to think about AI, this memorandum was all about how to deploy AI.

13:32It kind of reminds me of when Chuck Schumer and a group of senators released a policy sort of platform or framework earlier in the year. It was way more accelerationist than you might think. It was not really about safety. It was all about American leadership. That was very much the tone of this announcement as well. There was a clear recognition that AI leadership is a non-negotiable pillar of U.S. policy moving forward. Also, a clear understanding that a lack of clear policy has actually stifled adoption rather than allowing boundless experimentation. During his speech, Jake Sullivan warned,

14:13Frankly, this memorandum is remarkable in how clearly it sets a whole-of-government focus on a specific area of national security. Gregor O 'Allen, the director of AI at the Center for Strategic and International Studies, said, There are few precedents for a document such as this one which seeks to comprehensively state U.S. national security interests and strategy towards a transformative technology. NSC-68, which defined U.S. early nuclear strategy, comes to mind. Axios noted in a report that that report was among the most influential of the Cold War and opened the door to big spending on weapons and defense.

14:45And this certainly feels like a seminal moment with those types of implications. That, however, is going to do it for today's AI Daily Brief. Appreciate you listening or watching as always. And until next time, peace.