SB 1047: The Bill That Could Hinder AI Progress ft. Anjney Midha

0:00:06 Can you break down what is SB 1047? SB 1047 is a proposed law, but this one
0:00:08 is the most harmful to little tech.
0:00:18 When all your marketing team does is put out fires, they burn out. But with HubSpot,
0:00:24 they can achieve their best results without the stress. Tap into HubSpot’s collection of AI tools,
0:00:30 breeze, to pinpoint leads, capture attention, and access all your data in one place.
0:00:36 Keep your marketers cool and your campaign results hotter than ever. Visit hubspot.com/marketers to
0:00:47 learn more. Hey, welcome to the Next Wave Podcast. I’m Matt Wolf. I’m here with Nathan Lanz. And
0:00:53 today we’ve got a really important episode. Today, we’re talking with Anjanay Mida and he’s a general
0:01:00 partner over at A16Z. He was on the ground floor for companies like Mid Journey, and Luma, and
0:01:06 Anthropic, and some of the biggest AI companies in the world. And today, his message is really
0:01:12 important. He’s talking about legislation that they’re trying to pass right now in California
0:01:21 that could kill AI. And this bill is called SB 1047, and it will really, really hinder AI
0:01:26 progress if it gets passed. And in this conversation, we’re going to talk to him about what this bill
0:01:33 is, why you should care about this bill, some better options for AI regulation, and what we
0:01:38 can all do about it to make sure that the right regulations get passed and the wrong regulations
0:01:45 do not. So this is a fascinating conversation with a lot to learn. So let’s jump right in with
0:01:51 Anjanay Mida. Welcome to the show. Anjanay, thanks so much for joining us. We’re really excited
0:01:57 to talk to you about AI and AI regulation and AI investments. So thanks for joining us today and
0:02:02 how you doing? I’m doing great. Thanks for having me. Why don’t we start by getting a little bit of
0:02:08 background? Can you break down what is SB 1047, like in layman’s term? What do we need to know
0:02:14 about it? Yeah, so look, basically SB 1047 is a proposed law. It’s a California state bill
0:02:20 that’s making its way through the California legislature right now. That’s part of a much
0:02:26 broader wave of 750 or so new pieces of AI legislation that have been proposed in the US
0:02:35 since Biden signed the AI executive order late last year. But this one is the most
0:02:42 harmful to Little Tech. Little Tech is startups, open source researchers, academia. Unlike many
0:02:50 of those well-intentioned builds, which say, “Hey folks, AI is a powerful new useful technology,
0:02:56 and like many useful technologies like electricity or the internet, which has good and bad uses,
0:03:01 we should be thoughtful about how this technology is used. We should punish bad actors for doing
0:03:09 bad things with that neutral technology.” Unlike that rational approach, this bill is drafted to
0:03:15 attack underlying model researchers, scientists, and developers. And among other things, it’s trying
0:03:22 to place civil and criminal liabilities on developers of AI models, as opposed to focusing
0:03:29 on the malicious users of those models. So as proposed by this bill, overseeing these new laws
0:03:33 would be a frontier model division, which is kind of like a new DMV they want to form,
0:03:39 a new regulatory agency that would have the power to propose requirements on startups,
0:03:44 on researchers, on academia that would dictate if a researcher or an engineer could ultimately
0:03:50 be thrown in jail or not. Now, it’s so crazy that when this bill was proposed amongst tons and tons
0:03:55 of other bills, most people read it and said, “Okay, crazy bills like this get proposed all
0:04:02 the time. This is never going to get anywhere.” But the California Senate passed SB 1047 in May,
0:04:09 32 to 1. And so this bill is now slated for a California Assembly vote in August,
0:04:17 less than 60 days away. If passed, we are one signature from Gavin Newsom away from cementing
0:04:24 this into California law. And so this is an incredibly dangerous piece of well-intentioned,
0:04:32 but incredibly misguided regulation that is trying to make AI safer by focusing on the underlying
0:04:35 model instead of the malicious misuses, which is really where we should be focusing.
0:04:41 If that’s passed, I don’t see how you build an AI startup in California. Why would you take that
0:04:47 risk? You’d go to Texas or somewhere else. I mean, no rational AI researcher or scientist
0:04:53 is going to risk being thrown in jail just to pursue their research in California.
0:05:01 I think that if I had to be really sympathetic, I think he’s probably trying to elevate
0:05:09 the attention that AI gets, but it’s just proposed such misinformed drafting that’s
0:05:14 completely devoid from the way these models are actually researched, trained, and developed in
0:05:20 the real world. When it comes to frontier AI research, a local legislator who has no background
0:05:27 in AI development, technology development, and whose co-authors are frankly, I believe,
0:05:34 the most experience that the bill’s co-authors have is real-world experience in a lab that has
0:05:38 actually developed and productionized these models is a four-month internship at Google.
0:05:45 One of the co-authors of the bill is a well-meaning high school think tank. It is staffed by a
0:05:53 couple of high school researchers or something. From a substantive analysis, we’ve put out tons
0:05:59 and tons of critiques of the substantive pieces of the bill, but the process around this bill,
0:06:05 oh my god, I mean, it’s just a joke. That’s what ultimately led us to launching this website,
0:06:15 Stop SB 1047, two weeks ago, because after many attempts to provide the senator and esteem with
0:06:22 feedback on the problems with the bill and how to address it, and being ignored, our founders and
0:06:27 the Little Tech community just got super frustrated when every new revision of the bill just ignored
0:06:35 all that feedback and instead made the bill even worse. With this August vote deadline looming,
0:06:41 it’s just become so important and urgent to amplify those voices that are being ignored,
0:06:45 right? Like startups, researchers, the open-source community at large to voice their concerns.
0:06:53 We just wanted to amplify those concerns that Scott Wiener’s team has been ignoring. I have
0:06:56 learned now, even if you’re not interested in politics, politics takes an interest in you.
0:07:02 A lot of legislators, I think, especially in other states, are being quite thoughtful about
0:07:06 saying, “You know what? We’re open to feedback. Give us feedback,” and they’re making revisions to
0:07:11 the bill that actually address that feedback. That’s not the case here, right? This is a process
0:07:17 that’s been led by a legislator who keeps saying, “I’m open-minded to feedback.” It takes a bunch
0:07:21 of founders’ time and companies’ time, and then when you see the new draft, it addresses none of
0:07:27 the core issues. For a theoretical example of what this could mean, and you can definitely
0:07:35 correct me if I’m wrong, but let’s say there’s an open-source model out there that some people
0:07:40 developed and they put online, made it open-source. Somebody else grabs that open-source model,
0:07:44 uses it to hack into a government system. I don’t know. Something like that. They use the model
0:07:52 to do some bad actor stuff. The people who made the model are just as liable as the people who
0:07:58 actually did the hacking, right? That’s right. If you read the bill, what the bill is saying is,
0:08:05 if you open-source a model that meets some criteria that they put, which is completely arbitrary,
0:08:08 we can get to that in a second. But if you open-source a covered model,
0:08:14 you have to certify that this model cannot be used for any catastrophic harms,
0:08:20 and if somebody downstream picks up your model and does something bad with it, fine-tunes it,
0:08:23 changes it, modifies it in ways that you didn’t control and does something bad,
0:08:28 you are liable as the open-source developer for the harm they did. They’re placing this perjury
0:08:34 penalty on that developer, and you might go, “Okay, Anj, well, perjury is kind of,
0:08:41 that’s pretty severe. If you’re a guilty perjury, you get thrown in jail.” Yes, you do. What this
0:08:48 bill is proposing is that if an open-source model developer fails to certify appropriately, and
0:08:52 by the way, there’s no real definitions proposed yet. All they’re saying is this new agency will
0:08:56 have the full rights to determine these definitions in the future. You could potentially be hell liable.
0:09:00 I think that’s just crazy.
0:09:05 Yeah, definitely. So, is this only open-source, or does this apply to the closed-source models as
0:09:12 well? They’re proposing civil and criminal liabilities on all model developers, open-source,
0:09:20 closed-source. So, how is this disproportionately affecting the smaller businesses
0:09:26 building open-source than it is the open AIs, Microsofts, Googles of the world?
0:09:33 Oh, this is classic regressive tax, right? If you just think about a concept of regressive tax
0:09:38 versus a progressive tax, a regressive tax is something that disproportionately hits
0:09:45 less resourced people than people with more resources, right? And the way they’ve drafted
0:09:52 the bill by putting all of this burden of definitions that have no precise definition today,
0:09:58 what’s going to happen is, if this bill passes, this agency is going to get lobbied by Big Tech,
0:10:03 who has armies and armies of lawyers and compliance experts to shape the definitions
0:10:09 in their favor. And tiny startups, open-source researchers, academic labs, who don’t have all
0:10:14 those resources will just be left out in the cold. We’ve seen this happen with multiple industries,
0:10:16 and that’s what’s going to happen here as well.
0:10:20 So, it actually sort of helps some of these bigger companies with the regulatory capture that,
0:10:24 you know, they’re not outright saying they’re going for, but they’re probably going for, right?
0:10:30 100%. Let’s take one example from the bill, which the sponsor,
0:10:37 the bill sponsor, Scott Wiener, keeps saying, “Oh, look, my definition of what a covered model is
0:10:42 only applies to Big Tech companies because it only gets triggered by $100 million training
0:10:47 threshold.” Okay, well, hold on a second. The Big Tech company’s training budgets are in the
0:10:51 billion. So, first of all, if all you cared about was really just attacking and regulating Big Tech,
0:10:57 you would start your bill with the number B for billion, right? Number two, what even is a training
0:11:03 budget? There’s no such canonical definition today. This space is so early that, you know,
0:11:08 if I sampled the 16 different AI model startups that I’ve invested in over the last three years
0:11:12 for their definition of training, every single one has a slightly different meaning, right?
0:11:17 Pre-training versus post-training versus fine-tuning versus computing latent representations,
0:11:25 like past training runs. If I took Lama, if I’m a startup and I took Lama 3,
0:11:31 which costs, call it, you know, about $100 plus million to train, and then I fine-tuned it,
0:11:36 does their training expenditure apply to mine too? None of these definitions have been,
0:11:41 have at all been, the bill’s authors have proposed zero definitions around these
0:11:44 pretty important issues, right? Do you think that’s purposeful? Because, like, obviously,
0:11:48 if you leave it vague like that, that gives them so much power and control over all of this, right?
0:11:55 Look, I think there’s the generous interpretation and the, you know, the less generous one. The
0:11:59 generous one, you know, there’s this idea of Occam’s razor, right? The simplest explanation is usually
0:12:04 the right one. When I first read the bill, I was so worked up, I was like, wow, this has been
0:12:13 maliciously vague, right, to put this burden on model developers. When I then looked at the bill’s
0:12:18 authors and their backgrounds, then I realized that they just don’t know what they’re talking about,
0:12:25 right? I mean, I kid you not, there’s literally zero beyond, I think beyond one researcher on
0:12:32 that team who spent four months inside of a lab as an intern. They don’t have any experts on the
0:12:36 drafting team who’ve actually trained models, who’ve deployed them, who’ve worked at startups
0:12:40 for an extended amount of time that are frontier model companies. I mean, I just think they don’t,
0:12:43 I think they’re well-intentioned. I wish I could tell you they had the competence to have done
0:12:48 this maliciously. I think there’s good reason to believe they’re just way in over their heads with
0:12:54 no real-world experience here. Right, right. Don’t attribute the malice what you can, you know,
0:12:58 explain with ignorance or whatever. I don’t remember the exact quote, but that seems to be
0:13:05 what’s going on here. Right. We’ll be right back, but first, I want to tell you about another great
0:13:10 podcast you’re going to want to listen to. It’s called Science of Scaling, hosted by Marc Robárez,
0:13:16 and it’s brought to you by the HubSpot Podcast Network, the audio destination for business
0:13:21 professionals. Each week, host Marc Robárez, founding chief revenue officer at HubSpot,
0:13:25 senior lecturer at Harvard Business School, and co-founder of Stage 2 Capital,
0:13:30 sits down with the most successful sales leaders in tech to learn the secrets,
0:13:35 strategies, and tactics to scaling your company’s growth. He recently did a great episode called
0:13:40 How Do You Solve for a Siloed, Marketing, and Sales, and I personally learned a lot from it.
0:13:45 You’re going to want to check out the podcast, listen to Science of Scaling wherever you get your
0:13:57 podcasts. I’m curious. If you were an advisor to help with creating some regulation, are there
0:14:01 things that you believe should be regulated, or do you think it should just be open door,
0:14:07 let’s just push forward, accelerate at all costs, or are there some areas where you’re like,
0:14:14 okay, these are areas I think should be regulated? Oh, I’m absolutely in favor of regulation. Let’s
0:14:19 make it clear. Models are powerful tools like electricity, they can be used for good and bad,
0:14:26 and we should focus on preventing people from doing bad things with it. But this approach to
0:14:31 regulating the underlying technology and placing burdens on researchers instead of placing the
0:14:36 burdens on the misuses of the models is completely misguided. I have an issue with this particular
0:14:39 piece of legislation. I don’t have a problem with regulation, especially regulation that’s
0:14:46 thoughtful, that’s drafted in partnership with industry, that puts America first, that doesn’t
0:14:53 just hand away our entire AI startup industry to China. Yes, I’m absolutely in favor of regulation.
0:15:00 If you were asking me, if you were drafting legislation with policymakers to make
0:15:03 sure AI is developed safely and responsibly, what would you prioritize? I’d probably look
0:15:08 for three basic principles in that drafting. One, focus on the misuses, not the models.
0:15:13 Right? Focus on the malicious users, not the underlying infrastructure.
0:15:19 The second would be to prioritize concrete security problems over these sort of
0:15:26 super theoretical borderline sci-fi, doomsday terminator scenarios that they’re calling AI
0:15:32 safety. Those are not our most pressing safety issues where a model autonomously goes rogue
0:15:38 and launches a cyber attack on our power grid. That’s the plot line of a Schwarzenegger movie.
0:15:42 Right? What is happening, and I know this because our portfolio companies are being
0:15:46 attacked by this, we get approached by law enforcement agencies all the time, is in fact
0:15:53 good old-fashioned spearfishing, misinformation attacks, identity theft that are where these
0:16:00 attacks are increasing in speed and scale because bad actors are using AI tools. It’s the same attack
0:16:06 vectors. We have laws that say these are illegal. We don’t need more laws to say these should be
0:16:12 even more illegal. What we do need is laws to bolster enforcement, invest in defensive tools
0:16:16 that our agencies can then use to fight these increasing speed and scale of AI security. That’s
0:16:20 the problem we should be focusing on. Right? Anyway, that’s the second thing. Let’s prioritize
0:16:27 concrete AI security over sort of doomsday safety scenarios that have almost zero empirical evidence
0:16:33 that these will ever come to pass. Then I think the third thing I would do is to really prioritize
0:16:40 open-source development in the United States to maintain the competitive edge we have globally.
0:16:46 Right? Because us placing these burdens on our startups, our open-source researchers,
0:16:51 our universities is not slowing down China. They’re full steam ahead. But if you prevent our
0:16:55 open-source ecosystem from collaborating, from putting up models that people can research,
0:16:58 can fine-tune, can red team to make them more secure, you’re going to hurt us,
0:17:03 and you’re hurting the U.S. national of competitiveness. Nobody else, while everybody
0:17:08 else races ahead. Those are sort of the three simplest principles. We provided that feedback
0:17:15 ad nauseum, to be honest, to the senator, but none of the amendments to the bill have addressed
0:17:20 these core issues. Right, right. Speaking of China, I don’t know if you saw the news today,
0:17:26 but it looks like open AI is going to be banning chat GBT in China. It looks like this is possibly
0:17:29 in collaboration with the U.S. government, or at the direction of the U.S. government.
0:17:36 So I do wonder if we’re going to end up in a scenario where open AI and the other major AI
0:17:41 players who are closed-source, if they’re already in collaboration with the U.S. government behind
0:17:46 the scenes, a member of the NSA who joined is, I think, a board member of open AI recently,
0:17:52 Mira Moradi, the CTO, she openly said in an interview that they collaborate with the U.S.
0:17:56 government in terms of showing them the new models before they come out. I do wonder if
0:18:00 that’s going to lead to a world where, yeah, the closed-source models, they’re collaborating with
0:18:04 the U.S. government because the U.S. government sees this as a national security. It’s an asset
0:18:09 to the U.S., but also it’s a security threat. It’s a risk as well, and if they will actually end up
0:18:14 pushing that there should be no open source because of that. I’m almost kind of like what
0:18:18 Founders Fund is saying, where it’s like, and that’s one area where I’m a little bit conflicted,
0:18:22 because I saw some of the people at Founders Fund saying that open-source AI can be dangerous,
0:18:25 right? And that’s actually going to help China. So, I’d love to hear your thoughts on that,
0:18:29 like how kind of A16D is more on the side of open-source, and it seems like Founders
0:18:35 Fund is slightly against open-source for AI. Look, I think any arguments that claim
0:18:44 that open-source AI is a threat to national security are either, frankly, misinformed so that
0:18:50 they’re just coming from a place of not knowing the true state of reality on the ground
0:18:54 or they’re malicious, and that they’re designed to hold the United States back.
0:19:00 And let me explain what I mean there. Number one is a very, I think,
0:19:08 misinformed understanding of the state of information security at the best labs, right?
0:19:16 There’s this idea we have that closed-source labs are so protective and secretive of their
0:19:23 weights that China doesn’t have them, and we somehow have this amazing competitive advantage
0:19:30 over China. For over 10 years now, the Chinese government has had a state-sponsored program
0:19:37 to infiltrate targets of valuable IP development in the United States, and it’s not AI-specific,
0:19:44 this is in all kinds of industrial processes, that is a nationally-sponsored strategy by the
0:19:52 government of China to exfiltrate valuable IP from the United States to China. And while the FBI
0:19:57 and other enforcement agencies can’t comment on ongoing investigations, I will tell you that you
0:20:00 don’t have to look too far to look for public evidence that this is already happening at the
0:20:07 frontier labs. Just two months ago, there was an engineer from Google who was caught by the FBI
0:20:13 boarding a plane to China with TPU schematics on a thumb drive. We’re not talking sophisticated
0:20:23 exfiltration here, guys, thumb drive, okay? So, I would be, number one, I think any national
0:20:32 security game theory that folks are abiding on must take into account the reasonable likelihood
0:20:37 that frontier model labs in the United States are already infiltrated by adversarial nation states.
0:20:42 Frankly, I think there’s good evidence already from our enforcement agencies that ongoing
0:20:49 investigations that will soon become public will make that clear. But you just have to
0:20:55 go read the news to know that this is happening. So, number one, any national security strategy
0:21:01 that says, oh, we’re ahead and they can’t get our weights from closed-source labs is you’re already
0:21:06 giving away the game. Okay, so let’s start from an operating assumption that at best, we are at
0:21:12 par with them where they have our frontier developments today. I’m not even sure we can
0:21:19 claim we’re ahead. Let’s just say the goal is to remain at parity, right? The idea that open source
0:21:27 is somehow going to give away our national competitiveness fails to take into account
0:21:33 that the way we got to the frontier in the first place was through collaboration between researchers
0:21:40 of different labs, right? And the current big tech argument that, oh, open-sourcing our weights will
0:21:47 allow adversarial countries to get them does only one thing and one thing only. It allows them to
0:21:54 stop having to publish their research, stop having to have a convenient excuse to tell
0:21:58 their best researchers who want to, by the way, publish their research. The way the best AI researchers
0:22:09 get more sort of feedback on their research is by presenting openly. The scientific process
0:22:13 is you put out your research, you share about it publicly, other people then provide feedback,
0:22:20 and then you improve, right? That entire process of open collaboration at the frontier of AI is
0:22:27 about to basically be all but dead. And one of the biggest ways that we have shot ourselves in
0:22:33 the foot is by preventing academia in the United States to contribute to that research, right?
0:22:40 So if open source, for example, today is the only way that allows frontier university labs
0:22:47 to contribute to research at all. If Lama III was not open sourced, if Mistral was not open sourced,
0:22:56 Stanford, Berkeley, MIT, like these institutions, the postdocs, the PhDs there would have zero way
0:23:03 of contributing to AI research. And so I think if you believe that the public university system
0:23:08 and open collaboration between labs is critical to keeping our national competitiveness ahead,
0:23:15 then turning off open source is a great way to keep us behind, right? Especially at a time when
0:23:20 our labs are already infiltrated. So if the enemy already has our best and then we’re slowing down
0:23:26 our best, the most likely steady state is that we lose our national competitiveness and we fall
0:23:34 behind, right? So I think that either people don’t know just how people advocating for these who are
0:23:38 arguing that open source is bad for national security. Frankly, a lot of them just don’t know
0:23:42 what they’re talking about because I don’t think they’re investors in enough frontier labs at this
0:23:47 point. And frankly, I just think there’s a bunch of people who are culturally misguided because
0:23:54 they think that these doomsday scenarios are more realistic than they really are.
0:23:59 So I have one one sort of last question about the SB 1047. It’s a little bit of like a devil’s
0:24:06 advocate question. So if this is like a California bill, you know, when people just argue, well,
0:24:11 just go do your research outside of California. Like, I don’t know, just I’m curious your thoughts
0:24:19 on that. Yeah, it’s a good question. So unfortunately, this, the drafting of the bill was amended to be
0:24:26 even more clear that the bill stretches across state lines. You know, up until last week,
0:24:30 there was some debate like, oh, Ange, like it doesn’t say that this applies outside of California.
0:24:35 They, the bill’s authors went outside out of their way to make it clear that the statute would
0:24:44 reach across state borders. Oh my god. So really, I mean, I mean, I believe the bill’s authors has
0:24:50 been actually promoting that as a feature of the bill, not a buck. So this legislation is nationwide,
0:24:54 whether we like it or not, or they’re proposing it to be nationwide. So the only, I think the most
0:25:01 likely scenario will be that our best researchers, our best teams will move offshore to this emerging
0:25:08 kind of region across the world that I’m calling an AI sanctuary. Basically, there’s sort of three
0:25:15 things you need now as a world-class startup or a model research team. You need cheap electricity,
0:25:23 right, cheap, abundant, sort of sustainable, clean electricity to run the massive amounts of
0:25:27 compute you need to train these models. And the last thing is you need regulatory
0:25:32 certainty and protection to train these models where you’re not being as a researcher, you’re
0:25:38 not being held with civil and criminal liabilities. And you know what, frankly, I have been shocked
0:25:46 by how many, you know, nations have reached out since we started publicly speaking about the
0:25:51 bill saying, hey, please send us your best and brightest. We will gladly protect them without
0:25:57 regulations. And I think that will mean that our best companies do offshore to places that are
0:26:01 offering them cheap and abundant energy, compute and regulatory protection.
0:26:06 Yeah. So let’s talk about what people can do. If, you know, you’re listening to this and you’re
0:26:11 going, OK, yeah, this definitely sounds like we need to stop this from happening, what can we do
0:26:18 about it? Yeah, I’m glad you asked. So stop SB1047.com. It’s a public website that’s a hub for researchers,
0:26:22 academics, and anybody else concerned about the impact of the bill can go and write to their
0:26:28 legislators. So if you oppose the bill, please visit the site. We’ve got a templatized bill that
0:26:34 you can then customize for yourself and send it to your assembly representative. We have a list
0:26:39 there where you can easily pick who your representative is. We released the website
0:26:47 last week. And in the first four days, we had the community send over 375 letters to the assembly.
0:26:52 And so this is an important issue that a lot of people, a lot of startups, a lot of academics,
0:26:58 and a lot of open source researchers are concerned about. But we need to get the word out to even
0:27:03 more people. We have less than 60 days before the final assembly vote on this proposed law.
0:27:07 So please tell others about the site, share the information, raise awareness among those who’ll
0:27:12 be impacted by this bill. We basically think Little Tech deserves to have its voices heard.
0:27:15 And so if you visit the website, we make it super simple for you to understand
0:27:19 how this bill impacts you if you’re Little Tech and how to take action, which is to send a letter
0:27:26 to your representative and make your voice heard. You know, the message about helping small startups,
0:27:30 like I personally feel that, but I think a lot of people are probably going to resonate more with
0:27:34 the fact that this is very important for the future of America. It’s kind of appropriate
0:27:40 that this is like right after July 4th, and it can stay. We’re celebrating America and I personally,
0:27:45 that’s the thing that inspires me is like, this is very important. Like if we just like hand AI
0:27:50 dominance to China, like the reason America has been so successful is that we were, you know,
0:27:55 we were the leaders in culture for a long time with like entertainment. We were leaders in technology,
0:28:02 internet. And so that’s why freedom is spread around the world. And if we don’t win an AI,
0:28:04 it’s probably going to be the opposite of freedom that’s spreading around the world
0:28:10 through China. And so I think it’s a big problem. And like, and the opposite of freedom when you
0:28:16 have AI is, it can be quite scary when you can use AI to mass control people. And so I believe
0:28:21 that it’s really important that America wins this. And I think that more people probably will
0:28:24 resonate with that kind of message versus, you know, obviously, Silicon Valley people,
0:28:28 we like, yeah, we want to support the startups. Right. But a lot of other people, I think that’s
0:28:34 a more powerful message, I think. I think you’re absolutely right. And there’s a professor at
0:28:42 Berkeley, Ian Stoika, he testified in Sacramento last week saying that this bill, well, this bill
0:28:45 is called safe and secure innovation for frontier artificial intelligence models.
0:28:50 In its current form, it would do the opposite. It will hurt the innovation in California,
0:28:55 and it will result in a more dangerous rather than a safer world. And he goes on to explain that
0:29:02 first if SB 1047 passes, when it comes to open source models, he predicts that within one year,
0:29:08 we will all use open source models developed overseas, likely in China. Why? Because this law
0:29:13 will discourage building open source models in California, and likely the United States,
0:29:18 and Chinese open source models are already very competitive. And three of the top six
0:29:24 open source models are already Chinese, according to the Berkeley LMSIS chatbot arena evaluation.
0:29:29 The second is that if SB 1047 passes, then California will lose its competitive edge when it
0:29:35 comes to AI. Because as a researcher in a fast moving field, you don’t want to be constrained by
0:29:39 such limitations. So you just go elsewhere when you can do your best research. So more and more
0:29:44 PhD students of Chinese origins will just go back to China, while others might consider going to
0:29:51 places like, you know, other adversarial countries. So where they can enjoy huge funding for their
0:29:57 research. And this is already happening, according to him. And he’s a leading academic at one of the
0:30:02 preeminent American university labs. And so when he’s saying it, we really have to, I think,
0:30:08 sit up and pay attention. And then the last thing he did talk about is how SB 1047 incentivizes
0:30:13 companies that sell to enterprises to move out of California, since most of their customers and
0:30:18 enterprises have already have headquarters, you know, out of California. And so for the
0:30:22 California market, they could just provide, they will have to basically provide inferior models
0:30:27 to conform with SB 1047, which will mean that the state will will turn from a leader to a laggard.
0:30:33 And that’s not a future any of us wants. And so it’s this is, you’re right, Nathan, that this is
0:30:39 not just a California issue. This is an America issue. And I don’t think enough people across
0:30:44 the U.S. realize just how dangerous this piece of legislation is for all of America. And I think
0:30:49 more people should be talking about about about it the way you are. I think on that note, that’s
0:30:56 probably how we’ll wrap up the episode. But everybody can head over to stop SB 1047.com to
0:31:02 get more details about the bill as well as more details about how to help prevent this bill from
0:31:06 actually getting passed. And Anjanay, thank you so much for hanging out with us today. This has
0:31:10 been a fascinating discussion. And I think it’s really going to open a lot of people’s eyes. I
0:31:14 don’t think a lot of people even realize that this is kind of happening behind the scenes. It
0:31:18 doesn’t seem like it’s getting a lot of publicity right now. So I think it’s important that we
0:31:22 have these discussions and let people know that this is happening. This is what the California
0:31:27 government’s shooting for. So if you like what we’re getting out of AI right now, and you like
0:31:32 the progress we’ve seen, we need to do something about this. So I appreciate you sharing all your
0:31:36 thoughts and all the details about this, because I do think it’s going to be eye-opening to a lot of
0:31:41 people. Oh, thank you guys. I’m a huge fan of the bot and, you know, helping us spread the word and
0:31:48 get the message about the gauze out is deeply appreciated. So thank you. Absolutely. Amazing.
0:32:04 Thanks again. All right. Thanks, guys.
0:32:06 you