Inside ML@Scale #2

all the numbers behind ML@Scale (yeah, all!)

Every 25th of the month, I publish Inside ML@Scale — a full, honest look at the business behind this newsletter.

The good, the bad.

What’s growing, what’s stagnating, what I’m trying that isn’t working (or sometimes is?)

Why?

Three reasons.

1. Because nobody else does it.

The ML newsletter space is full of people performing success. “Just crossed 50k subscribers 🚀” with zero context on how, for how long, at what cost, with what retention.

I’ve been building ML@Scale since 2023: from 0 to 13k+ free subscribers, 45k on LinkedIn, and some of them paid.

That didn’t happen linearly. It happened with a lot of stupid decisions in between. Those are worth writing about.

2. Because transparency compounds.

The same way our 25th-of-the-month ritual makes us better at money, writing this publicly will make me better at running this business. You can’t hide from numbers you’ve committed to publishing.

3. Because you deserve to know what you’re paying for.

If you’re a premium subscriber, you’re not just getting “The Blind ML Review” or “The Zürich Feed”. You’re buying into a thing that either works or doesn’t. I owe you visibility into which one it is.

What’s in the free section (this part)

Every edition will have:

• Top-line numbers

• The one thing that worked that month

• The one thing that didn’t work

• What I’m trying next

Top-line: where we are in April 2026

• 13,000+ free subscribers

• 45,000+ LinkedIn followers

• ~25% open rate (still constant, need to work on this!)

• 6x/week LinkedIn cadence, 3x/week newsletter cadence

What worked this month

Lots of articles, both technical and non technical reonated!

The Barbell Market for ML Engineers

A Blueprint for Scaling Recommender Systems

THE ZÜRICH FEED [Edition #3]

Sub split:

• Barbell market: 46 free subs, 1 paid sub

• Blueprint scaling: 21 free subs

• Zurich feed: 43 free subs, 2 paid subs

What didn’t work

This followup on xAI RecSys is imho very high quality, but did not do as well as the first part. I expected more, I only got 22 free subs out of it.

xAI - Recommendation System deep dive [Part 2]

What I’m trying next

Focusing on a big month of X for august. Need to find time!

Also, I have unlocked drip campaigns on Substack, and I have setup them for:

1. New free subs

2. New paid subs

3. People that went from paid to free

Let’s see how they perform in the next month :)

What’s behind the paywall

Everything else. The detail. The numbers underneath the numbers.

• Full revenue breakdown by product (Blind ML Review, Zürich Feed, one-off paid posts)

• Conversion funnel math: where free subs come from, what % convert, at what rate, per channel

• Per-post revenue: which Lane 1 (technical) and Lane 2 (career strategy) posts drove money, which flopped, and why

• Churn: who’s leaving premium and when

• Eventually (when I get the courage) my personal financial situation alongside the biz. Same energy as the 25th-of-the-month ritual at home. Full picture or no picture.