Latest / Bad Decisions Podcast / Something Big is Coming, Anthropic's $30B raise, and Blender's DLSS
Transcript
- 0:00An article went viral on X with over 80 million views sounding
- 0:05the alarm about AI. Anthropic, the company behind
- 0:09Claude, raised $30 billion with a valuation of $380 billion.
- 0:14And do we finally have Blender real time viewport rendering
- 0:18with DLSS? We're going to find out.
- 0:21Welcome back everybody, to the Bad Decisions podcast.
- 0:23It's Friday and we are back. Yeah, and for anybody who just
- 0:27tuned in for the first time, this show is coming to you guys
- 0:30with three drops a week Mondays. Wednesdays and.
- 0:33Fridays where we give you guys everything unfiltered in the
- 0:37world of tech, AI and business. So if you want to stay up to
- 0:40date and get our analysis of what's happening in the world,
- 0:44this is where you're going to come to.
- 0:45And we also listen to your comments.
- 0:47We just did our first episode on Wednesday with over 50,000 views
- 0:52across all platforms, which is just insane.
- 0:54So we wanted to first start by saying thank you and we looked
- 0:58at your feedback, we look at your comments and we're going to
- 1:01be taking that into account to improve the show on that weekly
- 1:04basis. Are we ready?
- 1:06Let's get right into it. We are going to start with the
- 1:08article. So Pratt, if you can pull it up,
- 1:11this thing has been probably the most viral article I've seen.
- 1:15I've been seeing it for the past three days everywhere.
- 1:18Everyone is retweeting it, everyone is commenting about it,
- 1:21everyone is posting about it. Look at it now, 80 million
- 1:25reads. That's huge.
- 1:26I do want to say one thing because I also saw people on
- 1:29Instagram talking about this and we read through the entire
- 1:33thing. It's pretty lengthy.
- 1:35We're going to analyse it with you, but we do recommend that
- 1:38after the show, you go in and read the entire article yourself
- 1:40because at the end of the day, we are summarizing.
- 1:43This thing is pretty damn long. But Farah, what is it actually
- 1:46about? I think the the the way that the
- 1:49guy opened the article is really interesting.
- 1:51It's talking about how the current era is similar to how we
- 1:57were impacted during COVID times, how during COVID,
- 2:00everybody thought, oh, it's just a virus, nothing is going to
- 2:03happen, we're going to be fine. And then they were running
- 2:06around looking for toilet papers and work from home and all those
- 2:09Zoom calls. What is he trying to say is that
- 2:12in this AI era, the changes that we're going to face is probably
- 2:16going to be even more drastic than the COVID time and COVID.
- 2:19Was pretty damn drastic. I mean, we were locked at our
- 2:22homes, no one could go to school, university or work.
- 2:26And the way we all worked was substantially changed.
- 2:30And So what he's saying is that we're going to experience
- 2:32something even bigger with AI. And he broke it down to the Y
- 2:36and we're going to go over that. But before I begin, we have to
- 2:40give him some credibility or else you guys don't even want to
- 2:43listen. So this article is written by
- 2:45Matt Schumer. And for the past six years, he's
- 2:48been working on building an AI startup and also invests in the
- 2:53AI space, which means based on him, he is working with software
- 2:57and hardware and AI on a daily basis.
- 3:00He's building applications and he has experience, which is why
- 3:05when he experienced the disruption in the software
- 3:09engineering space, in the technology space, he decided
- 3:13that instead of having cocktail party conversations with people
- 3:17about AI, where he dumbs down what is actually going to
- 3:21happen, he wants to go ahead and write an article that is going
- 3:24to be understood by every single person in the world and that he
- 3:28gets to fully flesh out what he thinks is going to happen.
- 3:32I think he was saying that you can send this article to all of
- 3:35your family and friends, especially to people who don't
- 3:38understand AI, because it depends on where you actually
- 3:42live. Your exposure to AI or AI tools
- 3:44or AI news would be very different.
- 3:46Of course, if you're in LA or New York or if you're in states,
- 3:50your exposure would be much, much higher.
- 3:52So he wanted to give this article to give it to your
- 3:54family and friends, people who don't even know what AI is and
- 3:58they will get informed and make informed decision.
- 4:01So yeah, he straight up started by saying that he doesn't mean
- 4:05to be polite in this article because he wants to be real.
- 4:08As real as possible. And he first start talking about
- 4:12the intelligence explosion, right?
- 4:13The fact that the models are getting better by wider margins
- 4:19in shorter times, which means every single update that we're
- 4:23receiving with Claude, with Gemini, with ChatGPT and Grok,
- 4:27these new models are substantially smarter than the
- 4:31last ones. And the updates are coming in at
- 4:34a shorter gap. And that means we are seeing an
- 4:37exponential rise in the technology space rather than a
- 4:41linear 1. So he says if you think there's
- 4:45something that the AI is not good at now, think again,
- 4:47because these models are getting so good that at some point it's
- 4:51going to catch up, I think. Especially in the past few
- 4:53months that improvements have been insane.
- 4:57We got the ChatGPT codecs 5.3. We got Opus 4.6.
- 5:02These are revolutionary models that they can code for coding,
- 5:05Yes. And he was talking about how AI
- 5:08is building itself now. So they are using AI to improve
- 5:12the models and improve the language models that they have.
- 5:14And he's saying with AI doing the actual coding now, the speed
- 5:18would be dramatically faster. See, that's the most important
- 5:21thing. He even coded the Open AI team
- 5:24because once they released GPT 5.3 codecs and for anybody that
- 5:28doesn't know, that is the model that you use to write code with
- 5:33open AI software. And they mentioned themselves
- 5:36that it helped build itself. Now you have to understand what
- 5:40that means. If most of the grunt work of
- 5:44building the better the newer version of AI was done by
- 5:47itself, that means at some point it's going to be entirely
- 5:50automated. That means without human
- 5:52intervention, we're going to see open AI ChatGPT model get
- 5:56smarter. So Sam Altman and his team don't
- 5:59need to actually do any work. The model will make itself
- 6:03better. And now we are talking about AI
- 6:05companies. Spotify, I think, announced that
- 6:08their best developers haven't written a single line of code
- 6:12since December. Spotify is a huge company.
- 6:15They are servicing hundreds of thousands of people and since
- 6:18December they haven't wrote a line of code.
- 6:21They're simply reviewing code. That's what most software
- 6:25engineers who are in Silicon Valley or the big companies are
- 6:28saying now. They're not writing.
- 6:30They're just reviewing the code that the agents give to them.
- 6:33And I can only speak from our own example here at Bad Decision
- 6:36Studio. Let's talk about myself, Farhad,
- 6:40and Pratt, who is producing the show right now for you guys.
- 6:43All three of us do not know how to write code.
- 6:46However, since the past few weeks, we've all been building
- 6:50applications in house, every single one of us on their
- 6:53computers. Right now, I can look around and
- 6:55I see the monitors. We are building our own
- 6:57applications, so let's start. What are you building?
- 7:01So basically I'm building an app to distribute actually the
- 7:04content of this show. So with one click we can upload
- 7:07it on YouTube, Instagram, LinkedIn with captions and tag
- 7:11and everything integrated. And the captions and tags are
- 7:13created by the application Pratt is developing called
- 7:16Shakespeare. That's the name we've given it.
- 7:18And that application has been trained on all of our previous
- 7:21posts and all on all the top trending posts on different
- 7:24platforms. And it writes the caption
- 7:27contextual to the platform once we end this episode.
- 7:30So once we end this episode, right now, it's going to go to
- 7:33the AI and it's going to write everything out for us.
- 7:35I'm working on a dictation application.
- 7:38All of this has been done without us writing a single line
- 7:41of code. And this is precisely what Matt
- 7:44is talking about in the article. If software engineering is the
- 7:48first one that is getting disrupted, it's not going to be
- 7:51the only one. And he's trying to raise the
- 7:54alarm for it because he's trying to say that we didn't mean we
- 7:58didn't intend to come and replace software engineers.
- 8:01The only intention was if we're going to make AI better, we'll
- 8:05need to teach itself to make itself because making great AI
- 8:09models requires lots of code. So the first thing they tried to
- 8:12teach it was how to write its own code.
- 8:14And therefore we're at a point that we are today.
- 8:17So he's saying you have to now go look at finance, go look at
- 8:20law, go look at every other industry and think very deeply
- 8:24about how this could come and disrupt your industry so you can
- 8:28prepare better. Actually, he was talking about
- 8:31how the tech industry is impacted.
- 8:33Based off our own experience, we have seen the disruption in the
- 8:36creative industry too. Since November 2025, when we are
- 8:42going out, I have started seeing AI generated commercials,
- 8:46whether it's on billboards, whether it's an airport, whether
- 8:49on social media. This wasn't the case last year.
- 8:52This wasn't the case 24 months ago.
- 8:55So as much as the tech industry is getting impacted, the
- 8:57creative industry also get is getting impacted.
- 9:00The level and the type of companies that are using these
- 9:03tools are very different. And when we go to the anthropic
- 9:06section, I'm going to tell you how much companies are spending
- 9:09money using cloth. So imagine if a company is
- 9:12spending $1,000,000 on Claude API, just imagine how many line
- 9:16of codes they can create. That would be insane.
- 9:19One of the other things that I really like that he mentioned is
- 9:22when he addressed skepticism about AI quality.
- 9:26This happens in every industry. So first I'm going to refer to
- 9:29his example and then we're going to go to ours.
- 9:32He said that a lot of people's response to his article is that,
- 9:36oh, I tried ChatGPT and the answer was not good enough.
- 9:41And he was hallucinating when I asked it for something.
- 9:44And one thing he's saying is that a lot of the people that
- 9:46are saying that are using the free models of GPD, the free
- 9:49version of Gemini. And those are usually outdated
- 9:54models, models that only have knowledge up until a year ago.
- 9:58But if you go ahead and subscribe to the most premium
- 10:00versions, which is what we're doing, you're going to see a
- 10:03substantial difference in the quality that is outputting.
- 10:06And that is what's scaring him, essentially.
- 10:09Now coming back to our own industry, we come from the
- 10:12creative industry. The same is true when we talk to
- 10:15artists and we are telling them, hey, it's OK that you don't like
- 10:19what is happening now, but at least educate yourself on how
- 10:22the images are being generated, What models are working to
- 10:26create videos the best? Is it Clink?
- 10:29Is it VO3? Just get yourself updated
- 10:31because every single time we look at these videos and photos,
- 10:35we're seeing crazy amounts of difference in the way.
- 10:38In the last episodes we talked about seed dance, the amount of,
- 10:42you know, the, the realism quality and the amount of
- 10:44realism was really astonishing. And I think the the things are
- 10:48keep getting better. We are not going to stop here.
- 10:51This is not the end. The things are going to get
- 10:53better. And I think the successful
- 10:56people are people who know how to use these tools.
- 10:59So he actually gave some advice. He didn't want to be threatening
- 11:03only when he's writing this article.
- 11:05So he talked about the things that we can do to best prepare
- 11:08ourselves. Of course, one of them being the
- 11:11the ability to learn and to adapt because he said none of
- 11:15these tools will be the best forever right.
- 11:18You may be using cling 3 point O for your video generations
- 11:21tomorrow. Sea dance can come and be better
- 11:23today. You might be on Hicksfield as a
- 11:25platform. Tomorrow free pick is going to
- 11:26take over and the same is true for any other AI tool.
- 11:30So he said, learning how to be human again, learning how to be
- 11:34curious, right, learning how to be able to adapt so quickly,
- 11:38that is the most important skill that you're going to have.
- 11:42I think the next thing he mentioned was how you should use
- 11:45AI seriously, and I really like this topic that he covered that
- 11:49make sure you're using the premium version.
- 11:52So if you're someone who is getting a salary or who has
- 11:55client who's having an income, you guys should subscribe for
- 11:59for the most premium AI models, whether it's Open AI or
- 12:04Anthropic or Google Gemini or Grok, make sure you use the
- 12:08latest model because usually the free tier are not the smartest
- 12:12and they are one year or six months behind.
- 12:14And if you go to these different models, you can also see their
- 12:18differences because in here we of course use Grok, we use
- 12:22Anthropic, we use Open AI, and for certain tasks, certain
- 12:26language models perform better. So highly recommended to use the
- 12:30latest model available. And also you mentioned don't
- 12:32have any ego. I think that's also one of the
- 12:34biggest things. Even if you're a senior at a
- 12:37company or you've had years or decades of experience in your
- 12:41field and you just think that the AI will never be as good as
- 12:43you, That's probably the the biggest demographic that is
- 12:48going to get hurt. That's what he's referring to,
- 12:50that the people who refuse to acknowledge the rapid pace of AI
- 12:56are going to be hurt the most. So said, just have no ego, don't
- 12:59dismiss it. Just try to educate yourself and
- 13:01try to play around with all of the tools as much as possible,
- 13:05he said. One hour a day.
- 13:06We're spending pretty much the entire day, actually.
- 13:09Every day. What about you?
- 13:11I'm. Yeah, like every day.
- 13:13I don't remember a day that we didn't even use AII mean one
- 13:16hour. I don't.
- 13:17Understand one hour to educate yourself on AI.
- 13:20Education that's but I'm using codec so I have to.
- 13:24You're spending a lot. More time, so I have to educate
- 13:26myself, but I think it would be very important to you also know
- 13:30that this ego problem usually comes with when you have a lot
- 13:33of experience, when you have a lot of, you know, expertise in a
- 13:37certain craft and you don't believe that AI can come and
- 13:40disrupt it. So I think it's super important
- 13:42that even if you are super experienced, if you've been in
- 13:44the field for 20 years, still take a look at it.
- 13:48It might even help you do your work better.
- 13:50So check out the article for yourself, make sure you take a
- 13:54read and tell us what you think about in the comments, where you
- 13:57stand with all of this. Because Next up, we're still
- 14:00talking about AI, but we're going to a different style of
- 14:03news. What do we have?
- 14:04Here, no, we are going to cover Anthropic $30 billion raise guys
- 14:10$30 billion. That's that's a lot of money.
- 14:13What they mentioned on their official website that this $30
- 14:17billion going to be used for frontier research, product
- 14:21development and infrastructure expansion.
- 14:24I think most of it will go for power plants probably.
- 14:26It could. I mean, I'm not sure how they
- 14:28exactly will use the money, but what's insane here is that they
- 14:32have a 33180 billion dollar post money valuation and that means
- 14:37that they were valued at $350 billion dollars and after the
- 14:41investment of 30, they're now $380 billion valuation.
- 14:46This has of course given some critics on X the means to go out
- 14:51and talk about this deal and. Who's our number one critic
- 14:54Critic all the time? Elon.
- 14:56Of course, Elon. But Elon is criticizing
- 14:59Anthropic for a different reason.
- 15:00You want to talk about that? First, he's criticizing every
- 15:02single AI company though. That is not XAI.
- 15:05That is not. XAI So what is Elon?
- 15:08Saying no Anthropic posted on their X platform that they just
- 15:11raised 30 billion and Elon commented your AI hates whites
- 15:16and Asian especially Chinese and heterosexual and men.
- 15:19This is misanthropic and evil. Fix it.
- 15:22So for people who don't know anthropic means human and
- 15:26misanthropic, It means that is anti human or disliking in
- 15:31humans. So he is telling AI companies
- 15:33that your name is actually the reverse of your action.
- 15:36How he called open AI, closed AI and now anthropic misanthropic.
- 15:41So that's Ilan's criticism of it, but there's also other
- 15:45people who are criticizing the valuation.
- 15:48They're saying that this is overvalued because they try to
- 15:51compare it to something like Alphabet, the parent company of
- 15:54Google. And the current run rate of
- 15:57Anthropic is $14 billion. So that is their expected annual
- 16:02revenue, which has been growing 10X per year for the past three
- 16:07years. Wow.
- 16:08Which is insane, right? Which is insane, but just.
- 16:12Look at the. Yeah.
- 16:13So if you can, you zoom into that Pratt so you can see it a
- 16:17little bit more. Guys, January 2023, they were
- 16:21not making money. Yeah, $0.00 100,000,001 billion,
- 16:2514 billion. Yeah, and that's expected to
- 16:28grow considering again, this isn't a speculation, right?
- 16:31This is not a guarantee. If there are no changes, which
- 16:35we're always going to have changes, right?
- 16:36You're going to have a model that can be better than Claude.
- 16:39The reason why they're generating so much money, a lot
- 16:42of people believe is also because of Claude code, right?
- 16:45Because Claude code can write really well written code.
- 16:48But if tomorrow, you know, codecs just completely takes
- 16:52over or X I saw Elon saying with Grok that they're releasing much
- 16:57better coding models. They can definitely take a hit
- 17:00when it comes to the revenue. But again, the overvaluation is
- 17:03something that scared a lot of people because they were
- 17:05comparing it to alphabets public valuation, which is 3.7
- 17:10trillion. It's a lot, but they're saying
- 17:12something is mispriced here because if Anthropic, it's 380
- 17:15billion and Alphabet is 3.7 trillion, that's 10% of
- 17:20Alphabet. And Anthropic is technically an
- 17:24AI startup, whereas Alphabet has been a tech giant that has been
- 17:28matured for so many years. So they're saying this could be
- 17:32like the.com bubble. They're overvaluing everything
- 17:35and this is probably like what we experienced in the 2000s.
- 17:40But at the same time, some other people were saying, you know
- 17:42what, with AI, things are different now.
- 17:44Yes, much faster, right? It's different than before.
- 17:47So they're saying it's OK. I don't know what you guys think
- 17:49about it. Do you have any thoughts on?
- 17:51This no definitely. Anthropic is the leader in the
- 17:53category of enterprise AI, right?
- 17:56They are servicing all the big companies and some interesting
- 17:59stats. Two years ago, only a dozen
- 18:02customers spent over a million with them in a year.
- 18:05So a dozen customers spend $1,000,000 a year.
- 18:08Now they have more than 500 customer who spends more than $1
- 18:12million a year with them. So that shows again, the
- 18:15companies are using these APIs to write code, build product, do
- 18:19data analysis. And the moment you start using
- 18:22this model, it's very hard to stop because you you have
- 18:26probably replaced some of your workforce.
- 18:29Now they are doing coding, now they are doing development.
- 18:31They have their data with them as well.
- 18:32And then you need to keep developing.
- 18:34I think I I saw a stat where 4% of Github's.
- 18:39Yeah, I saw that. Code committed is from Claude,
- 18:43so that is a lot 4% of GitHub. From entire GitHub code.
- 18:47Yeah, that's a lot. And that number is expected to
- 18:49increase, which is why they're valuing it as this much.
- 18:52There's another, you know, claim that goes against the critics,
- 18:58which is this speculation is a private speculation.
- 19:01It's a venture fund speculation versus Alphabets, which is
- 19:06public because you then look look at the shares and the
- 19:10entire market. So this is something that people
- 19:13have been talking about on X, which is interesting.
- 19:15But you know for sure now looking at some of these news
- 19:18that most of the money in today's world is being spent on
- 19:22AI. And when you look at the news,
- 19:24the headlines are all AI related as well.
- 19:26They are getting a lot of investments.
- 19:28So part of the investors are Microsoft and NVIDIA, Yes, so.
- 19:32A huge part of the economy right now is focusing on AI.
- 19:36When we go to events here in Dubai or in Abu Dhabi, most of
- 19:41the booths have some sort of AI related activity happening.
- 19:45And the reason why this show even exists is because we're
- 19:48trying to keep up with all of the information.
- 19:50And so it goes back to the first article, right, that we covered.
- 19:54You do have to keep yourself up to date.
- 19:56If there's so much attention on one topic, there's a good reason
- 19:59why, and we urge you guys to also look into all of these.
- 20:02Things and and the great thing about all these tools is they
- 20:05are going to enable you to do things that you couldn't do
- 20:07before. As far as I explained a minute
- 20:11ago, we were not coders and now we are building apps.
- 20:15We are not building apps to sell or to distribute yet.
- 20:18We are building apps to solve our own problem within the
- 20:21studio. So if he wants something, he
- 20:23will make something for himself. If I want something, I can do
- 20:25it. Previously, I needed at least
- 20:28someone to look over the code, make sure the publishing is
- 20:31correct, make sure that our server and database is correct.
- 20:35But now we can do it. This is for us using it to do
- 20:38coding for us. Imagine an entrepreneur or a
- 20:41business owner or a freelancer, They might use it for their
- 20:44Facebook ads or meta ads to generate emails, to generate
- 20:47photos, to generate copies for their ads.
- 20:50So it really depends where you are and what industry you're in.
- 20:53But you can use all of these tools in order to improve your
- 20:57workflow. You know, the way I look at it
- 20:59is this, and this might be helpful to some of you guys.
- 21:01I'm tired of being pessimistic about the whole thing.
- 21:04You know, there are so many different pessimistic ideas you
- 21:08can find if you're looking for it on the Internet that will
- 21:11just ruin your day, especially when it comes to AI.
- 21:14But the way I like to look at it now is very simple, right?
- 21:18You can look at all the negative stuff, but what if you just look
- 21:21at it in this new perspective? What if you try to make yourself
- 21:26feel liberated with this technology, right?
- 21:28As Farhad mentioned, I couldn't go ahead and build any app that
- 21:32I wanted to before, and I can do that right now.
- 21:35For instance, we're using OBS to create this production for you
- 21:39right now. And Pratt has used some coding
- 21:43models like codecs and Cloudcode to create his own plugin that
- 21:47helps him create much quicker and much more efficiently.
- 21:50Something that we couldn't do before is not possible.
- 21:53So it makes you feel like you have a genie in a bottle that
- 21:56allows you to have wishes and it grants you those wishes.
- 22:00Now this is the interesting bit, right?
- 22:02I feel liberated because now I'm able to code even though I could
- 22:06never before. But I know for a software
- 22:09engineer looking at this right now, it's like, fuck, I used to
- 22:12be the one that would have coded this for them and they had to
- 22:15find me. Now they don't have to.
- 22:16But the same is true for the software engineer.
- 22:19The software engineer, if you wanted to make a website, he had
- 22:22to contact us for the, he had to ask us to make the images for
- 22:26him or the videos. Now he doesn't have to.
- 22:28Now he can use Nana Banana and cling and even clawed to create
- 22:32all of the visuals. So everybody at the same time is
- 22:36also more enabled. And so if you are tired of
- 22:40feeling all the negativity in the space, just sort of have
- 22:43this mindset in your mind. What couldn't you do before that
- 22:46you can do now with AI? And I promise you, it's going to
- 22:51open up a whole lot of doors for you as it has for us.
- 22:55I think you can agree with me on this, right?
- 22:57You've been blushing and smiling every day when you open up your
- 23:00computer. You're excited to come to the
- 23:02studio to code every day. No, it's funny because seven
- 23:06years ago I used to work in a company with a lot of software
- 23:09developers, and I had to work with a different team of
- 23:12software developers. Give the requirements, get the
- 23:14project back. Now I just work with codecs and
- 23:17it delivers. And you, it feels.
- 23:18It feels. Really.
- 23:19You're excited every day, Yeah, when you come into the studio
- 23:22because you're thinking of all the problems we have and the
- 23:25bottlenecks, and you're trying to solve it with code.
- 23:28And that's liberating. And that saves a lot of time,
- 23:30yeah. So that's why I'm saying it's
- 23:32very important to have that mindset.
- 23:35We're gonna move on to the next one.
- 23:37This one is pretty interesting. The Blender DLSS update.
- 23:41DLSS stands for deep learning super sampling, right?
- 23:44Yes, and they just recently announced it, but we've seen it
- 23:49in action. I think it was back in August in
- 23:52Vancouver during SIGGRAPH. Yes.
- 23:53So This Is Us in SIGGRAPH looking at Blender, DLSS at the
- 23:58Blender booth. So what you're looking at right
- 24:01now is the default denoising. So it's either optics or open
- 24:04image denoising, so by Intel and then we're going to switch to
- 24:07DLSS. Now you're going to see it real
- 24:09time moving around the scene with no lag whatsoever.
- 24:13So if you pay close attention now, this is this is the old one
- 24:18and then we have DLSS. Oh wait, what's the first one?
- 24:20DLSS. That was DLSS, So this is the
- 24:22old one, yes. So you switch it back.
- 24:24Yeah, there you go. That's real time now.
- 24:26So why is this so important? DLSS, for anyone that is
- 24:31unaware, is the AI powered rendering technology by NVIDIA
- 24:35that boosts performance and image quality on RTX graphics
- 24:39cards. So first of all, you need to
- 24:41have an NVIDIA card for this to work.
- 24:43And second of all, in this scenario of Blender, it's
- 24:46reducing the noise using artificial intelligence to boost
- 24:49the frame rate, right, and create the image for you at a
- 24:52higher resolution. It's also done in gaming, by the
- 24:55way. So in gaming, you can have DLSS
- 24:57on to be able to boost your frame rate.
- 24:59It's using AI to predict pixels, right?
- 25:01That's what DLSS does. But what's interesting here is
- 25:04that DLSS is a closed proprietary technology, while
- 25:10Blender is not. Blender is open source and that
- 25:14has caused a lot of pain for a lot of people in the community
- 25:17because when you're looking at a video like this, this is credits
- 25:21to Polyfiordi is an incredible Blender tutor on YouTube.
- 25:26So it's showing the differences of what DLSS can achieve.
- 25:29And it's just insane because in Blender we never had real time
- 25:34rendering. With Cycles you had to switch to
- 25:37EV which had far less realism to have a real time experience, but
- 25:42with Cycles you can now have that.
- 25:44It comes with a few limitations and he shows that and
- 25:47demonstrates it in the video. For instance, with fog, DLSS
- 25:51doesn't work that well. With reflection sometimes it
- 25:54doesn't work that well, so you have to be careful.
- 25:56But it's not going to be in the full build, right?
- 25:58It's not going to be part of Blender update.
- 26:01That is why I brought up the closed source and open source.
- 26:03SO if you go ahead and study the licensing of Nvidia's SDK and
- 26:10Blender, Blender is open source, meaning Blender's code is open
- 26:14to the public. You can look at it, you can
- 26:16study it, you can modify it, you can work on it, share it.
- 26:21And the rule is if you make the code better, you can go ahead
- 26:26and share it with the world. You can even sell things.
- 26:28But then you have to share your own code with others too.
- 26:31That's the whole point of open source.
- 26:33It allows for the entire market to create advancements in your
- 26:36project with Nvidia's SDK of DLSS.
- 26:40However, that is closed source, so.
- 26:42They cannot go into together. That's a problem with NVIDIA.
- 26:45They keep improving the technology, but for Blender to
- 26:48be able to launch this publicly and officially under a branch
- 26:53that they release this part of the code, the DLSS has to become
- 26:56open source. And we talked to the guys at
- 27:00Blender and Paula Fjord's video mentions that too.
- 27:03We even had a podcast with the founder of Blender that you can
- 27:05go ahead and watch where he talks about.
- 27:07He talked about vibe blenders open on how how long he fought
- 27:10for it. He fought for for such a long
- 27:12time, almost went bankrupt, but he got back and made it open.
- 27:16So it's just you cannot have DLSS launching publicly in an
- 27:20open source software unless NVIDIA decides to change their
- 27:24mind, which I doubt. Probably not going to happen.
- 27:26However, this is where it gets interesting.
- 27:28We have two other denoising methods in Blender right now.
- 27:31We have open image by Intel and then we have Optics.
- 27:34So knowing that DLSS is possible now and people have seen it, the
- 27:39entire community is not going to push for bringing a denoising
- 27:42method similar to that. That is going to be official.
- 27:46So you may have something like this in the future, maybe it's
- 27:48going to be the LSS, maybe it's not, but it's just exciting to
- 27:52see what you can have inside Blender in the future.
- 27:55A real time viewport that allows you to create so much more
- 27:58creatively. We do have that in Unreal Engine
- 28:00right now, but I just wanted to clarify that we do have that in
- 28:04Unreal Engine if you guys want to come and check it out.
- 28:06But jokes aside, Blender DLSS was cool.
- 28:10We tried it out. If you guys want to try it out,
- 28:13it is possible. There are.
- 28:15Tutorials for it. There are tutorials, Polyfiords
- 28:17video refers to one that allows you to compile and build your
- 28:21own version of Blender with DLSS in it so you can test it out.
- 28:25Just be aware of the limitations and the fact that it's not an
- 28:27official launch, so if you have issues with client projects.
- 28:30It's on you. No, no, no.
- 28:31Don't trust your client projects with that.
- 28:33Yeah, it's it's going to be on you.
- 28:35Well, I think that's it for today.
- 28:37Well, thank you very much everybody for tuning in.
- 28:39Remember we are coming back every Monday, Wednesday and
- 28:43Friday on X, Spotify and YouTube 1.
- 28:47Make sure you comment down the kind of things you want us to
- 28:49talk about, especially when it comes to the world's news in the
- 28:52tech space, AI space and business.
- 28:55We'll cover them in the next episode.
- 28:56And also, if you're watching this on any platform, just make
- 29:00sure you snap shot it, take a photo of it, share it with us on
- 29:03X or on Instagram. We'd love to see how you guys
- 29:05are watching us. And that's it for today.
- 29:08We'll see you guys on the next one.
- 29:10Ciao. Ciao.