Jtomm sass
I am less confident than I was a couple months ago when it was a dumpster fire, a productivity enhacing dumpster fire yes, but still a dumpster fire.
But I think my high level take is still this
software as capability ie image identifcation, language translation etc = screwed
software as a system level thing embedded across enterprises= safer.
I am still not convinced saas providers are all in trouble.
Enterprise infrastructure software is still hard to replace. its not just a thing that does X. there is years of data, its integrated into systems all over the company and it works.
It seems to me its much easier to embed AI flows into something that already works then it is to vibe code something and try to figure out all of the complicated ways it plugs into corporate infrastructure. To that end my sense is if you are an established player with a great product you get cheap AI coding too only you have more resources to stay ahead of newer competition.
And at least as of now Opus while really good still needs human guidance, and interaction to provide sane well designed code that is maintainable and extensible. I don’t see someone sitting down with the salesforce users guide and plopping out an identical product 4 hours later. That of course could change, and I’ll admit I’m surprised how far this model has come.
I have heard 3 fears1. companies are just going to vibe code workday. I am still very skeptical of that, and think it makes no sense
2. seat licenses no longer work. OK yes maybe, but billing structure seems solvable
3. new entrants. This looks more plausible than it did before simply because of the massive improvement in this one model.
I guess what I’ll say the threat seems overly discounted to me at this very moment.
I mean I could have recreated pinterest even before vibe coding . Instagram was a couple dudes in the basement IIRC.
If it’s as easy as vibe coding knock offs some of these should have already been supplanted.
AI
Market Data 
Markets 
