Serge Boupda

Quant researcher | Ex-SocGen exotic derivatives | African private credit strategies | Occasional thoughts on markets, finance and programmable economies

[Boris Cherny @bcherny]

As engineering, product, design, DS, etc. melt into a new kind of role, I was reflecting on what roles might look like in the future. For example, when I look at the Claude Code team I see what I think is five archetypes:

1. Prototyper: comes up with brand new ideas; churns out many ideas, most of which don't ship
2. Builder: quickly turns a prototype/idea into production-grade product/infra
3. Sweeper: cleans up the UI, simplifies the code and system, unships, optimizes performance
4. Grower: takes a product that has been built and iterates on it to improve Product-Market Fit
5. Maintainer: owns a mature system to make it secure, reliable, fast, and efficient as it scales

Many people span across 2 roles, and sometimes 3 roles. I also notice that these roles are not really tied to job function -- eg. across Anthropic, some designers match category 1, some 2, some 3; same for engineers, PM, DS.

A healthy team needs a mix of these, depending on the product:

- A product that is new and pre-PMF needs people that are strong at 1+2+3
- A product that is growing and has found PMF needs 2+3+4 and some 5
- A product that has strong PMF needs 3+4+5 and some 2

Maybe product roles of the future will look more like this, and less like the domain-specific roles of today?

[Serge Boupda @sboupda]

@daveg Cost of testing your idea: 📉
Cost of *distributing* your idea:📈

[Boring_Business @BoringBiz_]

This is one of the best primers that exist on the data center and AI industry right now

If you want to better understand the unit economics of each layer in the AI stack, I highly recommend you give this a listen

Chase Lochmiller, CEO and Co Founder of Crusoe, breaks down the inputs and outputs of data centers at a granular level

Shoutout to @apoorv03 for hosting yet another fantastic class

[Serge Boupda @sboupda]

@sytaylor Critical infrastructure work…

[Fence @fence_finance]

Today we announced Fence’s $20M Series A, led by @GalaxyHQ and joined by @paraficapital & @crane_vc

We’re using it to rebuild the operating infrastructure for asset-backed finance: the systems that verify assets, enforce facility rules, calculate obligations, and move capital.

[Serge Boupda @sboupda]

@eajene Dude… the label “continent” is reserved to its neighbor to the east (with quite similar issues and dynamics btw).
That being said, your point about scale makes the entire difference!

[Serge Boupda @sboupda]

@daveg Definitely… systematically add services that maximize float liquidity on the platform, and monetize through internal rates of conversion… and they have yet to monetize the banking and credit book curve the way a Nubank does…

[Serge Boupda @sboupda]

@biancoresearch Mark to model issue… looks like you’re essentially saying they should/should have hedge.d that risk.

[Serge Boupda @sboupda]

@BUGUR53177887 @Brad_Setser After selling gold outright…

[Serge Boupda @sboupda]

@michaelxpettis *addressed

[Serge Boupda @sboupda]

@eajene @theflutterwave The only way to monetize float reliably is build net interest spread and particularly, by exploiting cheap vs expensive convexity in a banking book…

[Serge Boupda @sboupda]

@Nicolas_Colin @daveg @nfergus @jamescrabtree @mgurri @profplum99 Nice. Welcome back to the noise factory, where gems are sometimes found 😉

[Serge Boupda @sboupda]

@mrstephendeng @alephile Here’s a controversial idea: traditional credit/insurance in Africa lacks of intermediation ie people moving risk in/out of their books bc of returns expectations or capital constraints.
I think that’s the bottleneck to solve

[Serge Boupda @sboupda]

@mrstephendeng Actually it’s a specific kind of payments data that is needed: properly tied to a purpose, filtered across request vs reply, labeled with a status, tied to “kyc’ed” users… otherwise it’s just white noise.

[Serge Boupda @sboupda]

@eoghan Great post!

[Eoghan McCabe @eoghan]

https://t.co/4bjaeYNcTP

[Serge Boupda @sboupda]

@eajene The ones with the big guns… 😏
More seriously, ratings don’t make the market price, bid/ask quotes do…

[Michael McNair @michaeljmcnair]

There is a real cyclical manufacturing recovery taking place.
People consistently miss the manufacturing turn, especially during large inventory cycles like we've had since COVID. The manufacturing PMI has been below 50 for 3 straight years. One of the longest manufacturing recessions on record.

To be clear, we have a structural re-industrialization trade AND a cyclical manufacturing recovery trade. These are different. Ex-China manufacturing has been getting run over by Chinese manufacturing, especially after China's credit surge to manufacturing designed to make up for weak FAI from property. But even with structural headwinds, manufacturing can have strong cyclical moves.

We've gone through an inventory whipsaw since COVID. Severe destocking, then double ordering and excess inventories, and now after 3 years of destocking, inventories throughout the supply chain have been worked down to levels we haven't seen since the GFC (outside of some areas with tariff-related hoarding).

Most people look at inventory to sales ratios, but that measure is misleading bc of the circularity problem. The sales number is impacted by the inventory cycle itself, which gets compounded the further down the supply chain you go.

I use an inventory measure created by Julien Garran @MacrostrategyP called the supply chain whip. The logic is to measure how much of a commodity is actually being used vs how much is being bought. The difference between actual usage and apparent consumption is the supply chain restocking or destocking. This lets you track inventory levels through individual supply chains rather than relying on aggregate ratios.

Inventories have become extremely low. Combine that with the strategic government stockpiling that is coming and you have the recipe for a sharp move.

When everyone destocks at the same time, it compounds down the value chain. Take this copper example from 2012:
China homebuilders see a 4% fall in sales, cut starts 13%. A 9% destock.
China AC manufacturers see a 14% fall in sales, cut production 24%. A further 10% destock.
Copper semifabs destock 700kt copper metal. Another 10% of Chinese copper demand gone.
Combined, copper saw a 29% destock along just the AC value chain in 2012.

People get surprised by manufacturing turns because they think in stocks when they should think in flows. One of the biggest ever inventory contributions to GDP came in early 2009 when the stock of inventory was still declining. But it was declining at a slower rate than it had been. After large destocking cycles, all it takes is less destocking to see a surge in orders. The second derivative.

To get a picture of how this works I'm going to borrow the wine cellar example from Julien Garran, who has been using it since his days as the Materials/resource analyst at UBS in the early 2010s.

Imagine you have a wine cellar (see the attached image):
Year 1. You have 100 bottles, drink 100, buy 100.
Year 2. Your consumption ticks up 10% to 110 bottles. But you think the trend might continue so you decide to hold more safety stock. You end up buying 130 bottles. That's 30% more than last year even though your actual consumption only grew 10%. Your cellar grows from 100 to 120 bottles.
Year 3. Consumption grows another 4.5% to 115 bottles, but you're happy with your 120 bottle inventory level. So you only buy 115 bottles. An 11.5% DECLINE in purchases despite consumption still growing.
Year 4. You decide 120 bottles is too much inventory. You drink 115 but only buy 95, drawing down your stock to 100. Your purchases fall 17.4% while your consumption is flat.
Year 5. You're done destocking. You drink 115 and buy 115. Your purchases surge 21.1% even though consumption is still completely flat.

The key point is that purchase volumes are far more volatile than consumption because they reflect both consumption AND inventory changes.

We're in a Year 5 situation now. People always get surprised by manufacturing improvement in these situations because end demand isn't accelerating and manufacturers are deeply pessimistic, right before they get a surprise surge in their orders. But you have to think in flows. Just LESS destocking leads to an improvement in orders.

This cyclical manufacturing recovery is driving the violent market rotation we're seeing...much more so than the Warsh nomination. Investors weren't positioned for it. They were completely offsides on this trade and have been forced to abruptly reposition for this new reality. Just look at the manufacturing stocks that haven't benefited from AI or caught a reshoring bid. These are the manufacturing industries where the US is long/already a large net exporter. Chemicals are the perfect example. After years of going nowhere, they've finally come alive.
The manufacturing cycle has turned.

[Emeka Ajene ✍🏽 @eajene]

"Discipline is doing what you hate to do, but doing it like you love it"

That quote from Mike Tyson really resonates—particularly the 'like you love it'

It's one thing to do painful work poorly

It's another thing to do it w/ vim & vigor

One checks boxes, the other compounds

[fintechjunkie @fintechjunkie]

https://t.co/9XHyFhQzBv