Moat of modelling in PE/Finance completely vanished;
I never believed it myself but I am vibe-modelling myself with Claude now.
I started college in ~2016 and I remember how hard it was back then to get really good at programming (e.g., statistics classes) and also financial modelling. Tons of hours of efforts went into mastering excel and all kinds of formulas. I used to study them from videos and take notes on different examples, etc.
All of that vanished. I recently saw an absolutely clueless intern (nepo girl) build a rather good three statement model on some subsidiary spin-offs. At the beginning I thought she is rather cracked, seeing all the formulas, but then I realized that she did nothing of this herself. Every datapull, every sumifs, every complex index-match, etc., all from the claude plugin. Even the three statement model itself with the raw data the client handed to us, everything was done by claude.
Kinda scary but I truly think modelling is just a commodity now.
// edit: anticipating questions. Right now I just 'vibe-modelled' bunch of annoying data gathering stuff. I.e., I had tons of input sheets and I wanted a few clean output tables.
Think of 10 different sub verticals/markets, 15 different regions, where partners want to have insight into each. It takes 2mins to let Claude build a table. How would we have done that 10 years ago? Tons of cleaning/manual gathering with Indirects, etc. - I think this would have taken at least 1h.
Always has been…it’s table stakes. Using a model to think commercially about a business / transaction is a different skillset, and something Claude can’t teach. My analysts do the same thing - semi-polished output on the surface, over engineered garbage in the inside. Can’t explain why things are built the way they are. And god forbid something busts and they wouldn’t even know where to look for the issue
Yeah we just tell folks, Claude can’t get fired but you can
Modelling has always been a commodity in PE. Nobody is promoted or paid because they are super good at modelling even for the VP promote. Super technical PE assoicates are very useful but never get the promote as that is the last rank being super technical has much or any value. You do need a base level of modelling skill to get the VP promote and to check over AI slop, which is why it is still a skillset you need to bulid. More valuable is being able to think about businesses, demonstrating an ability to make reasonable judgements / assumptions, showing that you are able to balance multiple stakeholders, etc. The higher up you go in PE the less technical skills matter, I am sure the billionaire founder of my fund has not opened a excel in decades at this point.
Sales skills is going to be where it's at for any FO departments that have been around for a while. Technical skills will matter moderately, but more and more technical work will be given to AI as capabilities advance.
Idk if I’d say all that. Truly thinking deeply about a business and what its place in the world related to every other factor to ever exist is still the primary generator of alpha. Sales has always been a derivative of the true value creation. PE gets away from that true value creation but that’s why it’s a 50 year industry that is going to complete and utter shit as we speak. There will be room for smart people in markets for like 15 more years until AI becomes god, I don’t think there will be a full shift to grifting until then.
i think AI will start taking away technical work in 1-2 years not 15, but I agree with your points that technical skills will still matter. Still it will be less work and more just understanding and interpreting and maybe opportunity identification.
There will be less need for analysts and junior level and instead one guy who's great at sales and understands technical/business side well can just be a one man army.
You'll see that judgement will become extremely scarce in the near future as more people delegate the thinking to AI. Judgement can be developed after thinking through lots of similar cases, not by reviewing how someone else thought about it.
Your value add to a firm has never been modeling. Claude makes it faster to do and lowers the already low barrier to entry of basic financial accounting and excel skills.
I think y'all think about this way too linear. I never said that the value add to a firm are purely modelling skills. Yet stil, the 'battle scars' of hardcore analytical work/thinking only get established if you spend 100's of hours deep into a model.
The anxious late nights where you tried to piece together why CAGRs don't add up -- why the EBIT bridge looks off and how price/volume/mix effects are somehow wrong.
I mean those are all endless of hours that basically sharpened critical quantitative thinking. There is no substitute for that.
I witness in the current generation that every request re model gets feed straight into claude. "Numbers don't add up here, can you provide an alternative"; I get that this is the Zeitgeist and the juniors are essentially forced to use it if they want to keep up speed. Yet, still, I already recognize that their thinking capablities drastically dwindle. They just don't get it. Simple questions, simple stuff I spot top down; they have absolutely no feel/smell for the numbers.
Totally agree - inefficient, yes, but having to build preposterous QoE bridges across multiple assets and currencies or doing some gnarly target leverage / acc-dil combo really taught me how to think outside the box in a way that you can’t prompt. And yes, seeing too much “why is the IRR too high” or “why doesn’t my balance sheet balance” type prompting without doing any of the work…
in 2 promts and 2 hours claude built a 3 statement model for a corp fi project that my dad fp&a md @ fortune 50 company said was far better and more in depth than anything his fldp's can do.
daddy
real mfs call their dads father
It's nearly impossible to collaborate on something when one person is using Claude for all of it. That's been my experience, and it's really frustrating.
But AI is tremendous to handle large data sets. That's where a lot of value is. It's so much better than us at it and a million times faster. Also, not for the more junior positions, but it's great support when going through legal docs (M&A, capital markets related, etc). Just my experience.
Multiplayer AI is coming and it's gonna be huge...
Well... yeah. The actual building of the model is a relatively low value part of the business.
It’s always some non technical TMT guy that’s model only has like 4 lines and a revenue growth input that makes these threads
Bump
yeah judgment is the moat now, nobody's getting hired for being fast in excel anymore
Quia ut neque beatae omnis est perspiciatis qui. Dolores vel non rerum et ut. Commodi sunt sequi et dicta dolorum.
Sit praesentium ea nihil laboriosam ipsam amet error dolor. Saepe consequatur accusamus voluptatibus eos officiis. Dicta rerum ducimus id asperiores dignissimos eos nam. Dolores molestiae omnis ut quasi nam ab incidunt. Quo nulla eum eum sunt debitis alias est. Possimus consequatur natus voluptatibus aut autem in.
Sunt sapiente maxime quae ipsam. Qui autem a molestiae temporibus. Tenetur ipsam est quasi asperiores voluptatibus atque. Voluptates omnis tempora blanditiis consectetur laborum et ducimus. Sapiente id aperiam dicta ducimus. Consequatur et mollitia est.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Necessitatibus adipisci odit sint. Rerum qui fugit a veritatis sit. Corrupti omnis nesciunt adipisci molestiae. Sequi aliquam quidem ullam sed.
Possimus ea ratione voluptate. Eius fuga consequatur autem eveniet ut quia. Voluptas eos distinctio ut illum. Omnis totam reiciendis velit doloremque ratione necessitatibus. Et ut aperiam eos et libero aperiam. Et dolorem et necessitatibus vitae dolore eos.
Ratione sint qui laboriosam commodi dolor. Qui est voluptate aspernatur asperiores et qui accusantium.
Laboriosam nam est laborum rerum id dolore tenetur expedita. Ut sed aliquam in. Corporis minus mollitia animi et ea natus et similique. Tempora eos sed minus quisquam.