Nathan Lambert

So in the US, you have all these platform SaaS companies, right? Salesforce and so on and so forth, right? In China, no one uses platform SaaS. everyone just builds their own stack because software engineering is much cheaper in China. Uh, partially because like people, STEM, number of STEM graduates, et cetera. Uh, so STEM is, so it's generally just cheaper to do. Um,

So in the US, you have all these platform SaaS companies, right? Salesforce and so on and so forth, right? In China, no one uses platform SaaS. everyone just builds their own stack because software engineering is much cheaper in China. Uh, partially because like people, STEM, number of STEM graduates, et cetera. Uh, so STEM is, so it's generally just cheaper to do. Um,

So in the US, you have all these platform SaaS companies, right? Salesforce and so on and so forth, right? In China, no one uses platform SaaS. everyone just builds their own stack because software engineering is much cheaper in China. Uh, partially because like people, STEM, number of STEM graduates, et cetera. Uh, so STEM is, so it's generally just cheaper to do. Um,

And so at the same time, code LLMs have been adopted much less in China because the cost of an engineer there is much lower. But what happens when every company can just invent their own business logic really cheaply and quickly? You stop using platform SaaS, you start building custom tailored solutions, you change them really quickly.

And so at the same time, code LLMs have been adopted much less in China because the cost of an engineer there is much lower. But what happens when every company can just invent their own business logic really cheaply and quickly? You stop using platform SaaS, you start building custom tailored solutions, you change them really quickly.

And so at the same time, code LLMs have been adopted much less in China because the cost of an engineer there is much lower. But what happens when every company can just invent their own business logic really cheaply and quickly? You stop using platform SaaS, you start building custom tailored solutions, you change them really quickly.

Now, all of a sudden, your business is a little bit more efficient too, potentially, because you're not dealing with the hell that is some random platform SaaS company stuff not working perfectly and having to adjust workflows or random business automation cases that aren't necessarily AI required. It's just logic that needs to be built that no one has built, right?

Now, all of a sudden, your business is a little bit more efficient too, potentially, because you're not dealing with the hell that is some random platform SaaS company stuff not working perfectly and having to adjust workflows or random business automation cases that aren't necessarily AI required. It's just logic that needs to be built that no one has built, right?

Now, all of a sudden, your business is a little bit more efficient too, potentially, because you're not dealing with the hell that is some random platform SaaS company stuff not working perfectly and having to adjust workflows or random business automation cases that aren't necessarily AI required. It's just logic that needs to be built that no one has built, right?

All of these things can go happen faster. And so I think software... And then the other domain is like industrial, chemical, mechanical engineers suck at coding, right? Just generally. And like their tools, like semiconductor engineers, their tools are 20 years old. All the tools run on XP, including ASML lithography tools run on Windows XP, right? It's like, you know, and like...

All of these things can go happen faster. And so I think software... And then the other domain is like industrial, chemical, mechanical engineers suck at coding, right? Just generally. And like their tools, like semiconductor engineers, their tools are 20 years old. All the tools run on XP, including ASML lithography tools run on Windows XP, right? It's like, you know, and like...

All of these things can go happen faster. And so I think software... And then the other domain is like industrial, chemical, mechanical engineers suck at coding, right? Just generally. And like their tools, like semiconductor engineers, their tools are 20 years old. All the tools run on XP, including ASML lithography tools run on Windows XP, right? It's like, you know, and like...

A lot of the analysis happens in Excel, right? Like, it's just like, guys, like you guys can move 20 years forward with all the data you have and gathered and like do a lot better. It's just you need the engineering skills for software engineering to be delivered to the actual domain expert engineer. So I think that's the area where I'm like super duper bullish of generally AI creating value.

A lot of the analysis happens in Excel, right? Like, it's just like, guys, like you guys can move 20 years forward with all the data you have and gathered and like do a lot better. It's just you need the engineering skills for software engineering to be delivered to the actual domain expert engineer. So I think that's the area where I'm like super duper bullish of generally AI creating value.

A lot of the analysis happens in Excel, right? Like, it's just like, guys, like you guys can move 20 years forward with all the data you have and gathered and like do a lot better. It's just you need the engineering skills for software engineering to be delivered to the actual domain expert engineer. So I think that's the area where I'm like super duper bullish of generally AI creating value.

Thinking larger than the context length.

Thinking larger than the context length.

Thinking larger than the context length.

I think it's that and then becoming a domain expert in something. Sure, yeah. Seriously, if you go look at aerospace or semiconductors or chemical engineering, everyone is using really crappy platforms, really old software. The job of a data scientist is a joke in many cases. In many cases, it's very real, but it's like,

I think it's that and then becoming a domain expert in something. Sure, yeah. Seriously, if you go look at aerospace or semiconductors or chemical engineering, everyone is using really crappy platforms, really old software. The job of a data scientist is a joke in many cases. In many cases, it's very real, but it's like,