Gil Luria

would install OpenClaw, but there's many companies that already have other instances of Claude and other uses of Claude that this is now a natural extension for.

So it does take it to another level.

It's a range.

I would say that, first and foremost, the companies that provide infrastructure software are the ones that appear to be more secure.

So security software, infrastructure software like Snowflake, Datadog, Microsoft...

are probably more benefiting from any growth in AI because you need infrastructure in order to deliver AI.

And so those are more insulated and more positively impacted by AI.

Then there's a whole range of companies that are probably more secure, companies that control a large part of the enterprise data schema.

how data is organized, and so then you're talking about your Palantirs, your ServiceNow, even your Salesforce and Adobe and Oracle to some extent, those are a little safer.

The ones that have had the most concern and probably justifiably so are companies that deliver

either customer center software or, again, workflow software.

Those are the companies that are most exposed.

Your Nice, your Five9, your UiPath.

Those are the companies that are most at risk.

And this just exacerbates the risk.

But again, the reaction is so strong that you have to step back and say, we are going to be using software, humans,

are going to be using software for a very long time.

And as long as that's the case, you need the same software, even if agents will be using the software as well.

The analogy to me is a lot like the internet.

Just because ChatGPT can go shopping for me online doesn't mean I don't also want to go shopping on these human websites as well.