A message from Octus: J.Crew, Xerox and Optimum walk into a restaurant...

LMEs aren't an off-menu option anymore. As balance sheets get squeezed and creative counsel finds new gaps in credit docs, the tactics are evolving faster than most investors can track. Uptier. Extend and exchange. Deal-away. Non-sub dropdown. Each one a different recipe. Each one with a different flavor of pain for someone in the capital structure.

Octus just published something worth bookmarking: The Stressed Chef, our LME menu breaking down the major liability management structures being used.

We mapped the transactions that defined these tactics, including old staples like J.Crew and At Home and newer creations like Better Health, Xerox and Optimum Communications. For each one, you get the base mechanic, the type of subordination being deployed and a spice rating so you can calibrate how aggressive the playbook really is.

The menu also includes a build-your-own framework so you can map any new transaction against its component parts: type of non-pro-rata purchase, subordination method and available extras like vote rigging, transfer restrictions and multi-step mechanics.

If tracking LME risk or advising clients on stressed credits is on your mind, grab a break with your favorite Chipotle bowl and our menu. 

Welcome back. Obviously, the advancements in Claude in mid-December 2025 really put the “AI is a bubble” thesis to bed, drawing into question software terminal values and rocketing the valuations of Anthropic and OpenAI.

The genie is out of the bottle with AI - it’s here whether you like it or not, so we need to talk about how finance professionals should be planning their careers around it.

Given how rapidly things are changing, we need to take a look at and examine the typical career paths and talk about what happens now. We’ll cover the following:

Let’s talk about these routes:

My overarching belief is that many finance institutions will simply need fewer headcount, as opposed to needing “zero” headcount. It’s important to remember that we’re shifting away from a process designed for Analysts/Associates and humans to find/extract data and more about creating an environment that amenable for agents. Are data and resources set up a way so an agent can go out and autonomously do xyz? The setup in the future is about enabling agents to get after it from a problem solving and end product creation standpoint.

For the Buyside, the overarching theme I’m going to talk about in this newsletter is how being someone who makes investment decisions or raises capital is a way to insulate yourself from AI. “Sales” is a job many of us do every day without knowing, and that’s something that continues to have material value. Ultimately, for Buyside roles and Banking classes, I think the class sizes are going to be smaller, but they will not go away entirely. So there will be the same amount of people competing for fewer seats. Banks and Managers are both consolidating too. Many of people have told me a lot of their class sizes are unchanged, so it’s not something we’re seeing in full force quite yet. I would expect a continued lag given many of these entry level roles are planned 18-24 months in advance.

Seperately, I have a harder time with the “Oh I didn’t hire an Analyst, I used AI instead” anecdotes. In many instances, it seems like it’s coming from a group where an Analyst or Associate hire is hard to rationalize. Sometimes, AUM, deal flow, and other rationale constrains your hiring ability. Full-on replacement is something I do not believe will happen at scale. The more probable answer is “I hired two Analysts/Associates instead of 3”. Let me get into it.

Banking - I have an extremely hard time believing that Investment Banking and lending money is “going away” - I would not have fear over that. The bigger question is how large the classes are and what roles get eroded. Clearly, equity and credit research isn’t what it used to be. Those seats are meant to be 2-3 year roles. However, with strong M&A and financing markets activity this year, the lure of being a Banker has become dramatically more attractive. Banking is obviously more cyclical than the Buyside (has fees, AUM atleast) but the seats closest to revenue and relationships are the most insulated. That’s obviously a big part of banking and buyside roles.

The workflow of a junior is going to change; much of what I did as an Analyst is now doable or will be replicated by AI. But a junior resource tied to grunt work, the mid-level professional that is in-training and reviewing the final product, and most importantly, the rainmaker senior professional all have a degree of insulation and a role to play.

The Buyside - Again, seats are going lower and there’s going to be some level of consolidation, but this is an integral role. The demise of private credit is greatly overexaggerated, and I think IG-rated private credit is a high TAM market. There are going to be some winners and losers around software though. Despite the recent software recovery, pricing power/switching cost and terminal value questions are very, very real. Software will get quite bifurcated. I don’t really buy into “AI will make everyone smarter so there will be less alpha” thesis. I still think deploying capital, and identifying public equity opportunities will still be skill-based and many of these funds are not going to let the most crucial final steps become agentic.

The buyside is going to get smaller, but private equity, private credit, CLOs, asset managers, LPs, and hedge funds are not going away. Like my thoughts around banking, AI is going to eat a lot of the busy work, plus memo creation and modeling, but it’s meant to free you up for the higher value tasks. One thing I worry about is while AI can do expert calls & run through earnings, I would argue you need a human element to determine fact from fiction. Because management teams lie, overexaggerate, or give social cues. Likewise, for expert network folks who sometimes overembellish things or are clueless about certain things. That human touch, the decisiveness, conviction, and in some ways “the gut” isn’t something that is necessarily replicable in real-world investment decision-making moments. A lot of the inputs and processes should absolutely be AI powered, but like my Banking comments, I don’t think we’re getting an AI Private Credit VP or an AI Pod Analyst.

AI Company/Startup route - there’s going to be a lot of opportunity for AI-enabled companies to succeed, but you need to ask the following: can Claude or OpenAI just do your business?

I think unfortunately a lot of vertical-specific AI companies are going to run into trouble when/if Claude or OpenAI just becomes the go-to one-stop-shop solution for corporations, investment managers, lawyers, engineers, and many others. Right now, it’s very much a 3-horse race with Anthropic and OpenAI racing against each other, and Google continuing to make strides with Gemini and capture a large part of the overall AI ecosystem. But with token costs rising and the bifurcation of high-value AI work vs. low-value AI work coming into focus, it seems like being model agnostic is coming more into focus.

You have to think long and hard about your value prop with AI. To me, the bubble in AI isn’t in Anthropic or OpenAI, the bubble is in a lot of the Series A/Series F stories that are becoming headscratchers to non-silicon valley folks. When we are on to rounds beyond a Series E, I think many folks wonder wtf you haven’t gone public or sold yet.

SMBs - I think the SMB and real business route is much more attractive than it was a few years ago. Even with all the competition from everyone trying to buy their own business. There are real, tangible businesses that can cash flow $250k-$1mm/year that are going to have staying power and durability. The problem is there’s so much search fund, self-funded, and LMM PE attention towards this space that the space is getting insanely competitive. Still, there’s an insane amount of real-world businesses that run into succession issues where buyers would very much happily complete an asset-sale to get some money for their hard work. The important thing to remember here is you need to buy a Company, not a Job. Sure, it will probably feel like buying a job for a set period of time, but if you’re unable to grow the business into something that operates without you, then you’re not creating actual enterprise value or something that can give you a real exit.

From a SMB path, you want to optimize for both 1) real ability to take dividends or earn outsized compensation in a way that doesn’t materially hurt the business and 2) have the ability to sell at an SDE or EBITDA multiple that makes sense and clears you at least $1mm post-tax, if not more. You also want to enjoy the work of course and have the competence to actually operate.

Many of these real world businesses are probably going to be untouched by AI, while a lot of software, tech-enabled, and knowledge based businesses don’t have attractiveness in the future. As someone in digital media/software, that’s what keeps me up at night. In another life, I would’ve loved to go the Small business buy and grow route, and that might be on the cards at some point.

Corporate roles - It’s starting to become clear to me that unless you have tech forward leaders, the average Fortune 500, or even a company with 750-4,000 corporate employees, is going to move slower with AI than many of us would like. They may struggle with implementation, they may have poor practices regarding who and who shouldn’t be tokenmaxxing, and there may be a level of bureaucracy or lack of motivation, to really push things forward. I’m a firm believer that even before AI you could lay off 15% of a corporate workforce and not notice a difference. I would imagine AI only amplifies this, but I don’t think it will immediately result in mass layoffs. So there’s some level of durability and attractiveness to work your way into a corporate seat. Many of the finance people I know who make the move outperform quite easily because they’re used to working crazy hours and now their job is “easy”, plus they received excellent training in an investment banking or adjacent type of role.

While compensation might be lower, there’s clearer longevity, where there’s a much clearer path to keep working towards 60-70. Many finance roles are higher earnings, but have fewer seats the older you get/can be quite cruel to later on your career. I think this matters more when you leave NYC or other finance hub cities. Being a mid-level professional at a corporation in a less-finance centric city like Phoenix, Minneapolis, or Houston might give you more flexibility than a mid-level investment professional in a less-finance centric city. Obviously, if you’re going the Corporate route, you need to think about the durability of your Company in an AI-centric world. You may not want to work at a Software company, but you probably want to work corporate at a HALO (Hard Assets, Low Obsolescence) like company. Corporations = slower to adapt AI = but safer career.

Banking roles - Outside of the core, front office roles, AI productivity gains probably leads to further rationalization in less core revenue producing roles. The bigger problem I think is going to be middle and back office functions that have workflows that AI can easily do faster. There’s still going to need to be human elements, it’s not like compliance is going to be all robots, but there’s going to be some level of automation added. Still, there’s massive regulatory burdens for a lot of these financial institutions, so over-applying AI in mission-critical functions could be more dangerous than it is helpful. But it’s important to remember that a lot of the issues and problems we’re talking about with AI today aren’t going to be problems in the future. What AI can do and associated hallucinations have changed dramatically in the past 6 months, 12 months, 18 months, etc. So you have to remember that we’re imagining a world where AI executes flawlessly, and what careers look like as a result.

Financial Advisor - I bring up the financial advisor route because it’s a role someone might evaluate later on in life. I think some advisors have a rep of actually lacking domain knowledge. I’ve found that a lot of the people I graduated with who became advisors were people I would never take advice from and have very surface level views of investments or markets. I hope I’m not insulting too many people but it’s what I’ve seen at the younger, learning your trade level. That’s a little different from someone who pivots into it later on in life.

We just cited a Bloomberg article in the Wall Street Rollup highlighting how wealth management jobs are now under pressure.

Based on the chart here, it seems like the financial advisor business might have a curve similar to the decline in linear TV/broadcasting:

I think a lot of the high-priced/high-fee “advice” economy is in trouble with AI, but maybe I have too much of a tech forward view. I’m more negative on the need for a financial advisor than the avg. person, so maybe you can ignore me.

Aside from the career path dynamics, I want to talk about training yourself to make sure you aren’t susceptible to losing key critical thinking and reasoning capabilities:

Avoid AI brain drain and social media brainrot like your career depends on it. Because it really does. I wasn't saying this advice in 2022, but with the rise of AI, it is either a tool that propels you or a crutch that cripples your ability to do the basics. Overreliance and the inability to think critically & build models on your own (just to make sure you can) will kneecap you. You need to know the ins & outs and know how to reach original opinions before letting AI boost you.

On Social Media brainrot: Yes, this is coming from a guy who posts a 12+ times a day. I take some level of comfort in being addicted to X, because at least I’m scrolling around getting sophisticated takes on individual stocks and the latest in AI. Instagram seems insanely more destructive, as does TikTok, which I’m not actively on.

The Instagram reels page is a nightmare loop, you should actually avoid pressing that button, it is designed to just entrap you. It genuinely feels like Instagram is working to make you actively dumber and turn you into an Idiocracy movie character. I appreciate how thoughtful a lot of the ppl who follow me & message me on Instagram are, but this financial markets niche is far from the reality of what Instagram has become. My account was very small in 2020-2021 but I’ve seen how Instagram has changed from 2020 to 2026, and it has gotten so dumbed down it’s not even funny. I have genuine doubts about the ability of the average Instagram user to read.

Last week, I satirically made fun of “firing” my intern (Drake Maye) and the comment section response was overwhelmingly negative. The sheer amount of people who thought Drake Maye, the Quarterback of the New England Patriots, was an Intern of mine who acted highly unethically was alarmingly high.

Even if you didn’t know who Drake Maye was (good thing Google exists), you’d still probably be sharp enough to realize a satirical finance page probably has a layer to the joke you didn’t initially get. This showcases the social media brain drain front and center.

You can set time limits on your phone, but you can also use external apps to create time limits for Instagram usage. I would likely recommend this for Instagram or TikTok unless you can honestly tell yourself that you’re learning a ton from the platform or have strong time-management where you’re getting your fix of entertainment but not getting stuck in an endless scroll.

Human edge: Honestly, this is really hard to convey. As a late 2010s professional, it feels like a blessing to be able to experience what working was like before COVID happened. A lot of people never experienced the pre-covid and pre-AI workforce.

However, I still feel like I was in the first wave of younger professionals addicted to their phone. And there’s definitely been a small level of stunted professional growth during days when I’ve had a hybrid schedule as opposed to 4/5 days in the office. But the reality is I took advantage of that to pivot into a career/path where I can, in many ways, do whatever I want.

There’s definitely a battle to stay focused during your workday though. I think one of the ways to overcome this is to go about your day with a level of purpose and production that makes you effectively and thoughtfully pound out the work that you need to get done. I love putting lists of what I need to get done, organizing it by priority/due date, and making sure if I’m still online at 10pm then there’s something quick I’ve been putting off that I can knock off before I go stream an HBO show.

If you talk to a Gen Alpha or a Gen Z student/intern born in 2004-2009 then you’ll realize that for many ChatGPT is just a tool to “cheat” or “study” with as opposed to turbocharging them. It’s not like these tools didn’t exist before AI though. Everyone used Chegg when I went to college. And everyone used SparkNotes when I went to High School. So that gives me some hope that maybe a lot of people are still able to learn by themselves, even with AI usage at record highs.

It seems like it’s going to come down to intellectual curiosity and whether you use the tools at your disposal in the wrong ways or as a superpower. Still, it’s not like you can bring your AI to an in-person test, so heavily weighted in-person tests are one way to incentivize students to learn. Some other advice is to do things at work the first few times in a way that you understand it, and then use AI once you get to the strong repetition stage.

Let’s look at 1) modeling and 2) writing CIMS/Memos + critical thinking:

The answer to getting skill-drained by AI is to simply go through the reps by yourself in an intelligent way. It’s about really understanding the nuts and bults and knowing why A is connected to B, what toggles what, and the overarching mechanics for modeling. I’ve partnered with Shortcut, an AI-Agent Excel model, and if you give it strong prompts, it pumps out the modeling that I used to do as an Analyst/Associate. So if you’re leveraging Shortcut you still want to understand the nuts and bolts of what makes a model a model and then start hypercharging your work product with AI.

Frankly, the short answer is you just can’t cut corners, and you need to put the work in. If you do that to make sure you’re nailing the foundation like the guys in 2015 did, then everything should work out fine.

The next step is building research memos. What was a no-brainer experience and thing you had to do in finance is now something that needs to be looked at in a much more raw fashion so younger professionals can truly understand what they’re doing.

Raw, Brute Force - My path was on Wall Street was very brute force & learn from mistakes oriented. I’ve seen people complain in comment section “I applied for 200 jobs and got nothing back, the job market is broken.” While just directly applying is a way a lot of people have gotten jobs historically (and still do) - the whole point of competitive job searches are to figure out as many ways to get your foot in the door as possible. That means via on-campus relationships with companies, speaking to an alumni or another connection, reaching out directly, networking, and other polite/professional, but in your face ways to have people keep you in mind. With the rise of AI, plus mass applying to jobs, fortune is going to continue to favor the people who can get preferential treatment in job searches. I’m sympathetic to it being a harder job market, but you can’t just go through the motions and expect your odds of getting in the room to increase. But you also need to make sure you’re ready once you get into the room. Many times, I got into rooms but slipped up because not enough technical chops or because I didn’t develop a granular or polished enough view. So you really have to marry brute forcing yourself in the door with technical aptitude. One of the things we’re building on Buyside Hub (my free job board, please sign up) is making sure the people with the brute force and aptitude are getting thrown in front of employers and finding the right opportunities.

If you want to check out my career resources - you can take a look at HYH Premium as well as the Banking Playbook.

HYH Premium is the credit career resources I developed to cover everything I needed to recruit for and land private and public roles. The focus is on zoning in on the exact things needed to thoughtfully put together an IC memo and drive modeling assumptions.

Meanwhile, Banking Playbook is a 248 page developed in collaboration with top Investment Bankers to help students prep. The full package is the best way to make sure you’re covering the networking, technicals, interviewing, and return offer essentials - you can learn more about it here.

While AI is a super power, you can’t really ChatGPT your way to landing a job, so make sure you’re leveraging the right paid resources and free paid resources in a way that gets you career ready.

Concluding remarks - This isn’t meant to scare you, but it is meant to level set in the world as it exists today and as it will change in the next 3-5 years. If you believe AI is a “bubble” still, I genuinely can’t help you. Nvidia and Google are at $5 Trillion valuations for a reason. OpenAI and Anthropic are already $900 Billion companies for a reason. The genie is out of the bottle. You have to deal with the hand that’s being dealt and act accordingly.

Like, my good friend, The Deep said, it’s a doggy dog world.

Until next time.