voidsnax

Open web display inventory is a mess. Nearly entirely fraudulent. Every person in the supply chain for open web display inventory, sometimes even including the Chief Marketing Officer of the firm purchasing the ads, is incentivized to turn a blind eye to fraudulent impressions. Fraud is ultimately paid for by shareholders and owners.

This is because everyone in the supply chain is paid per impression, so they don't have any real incentive to root out (or even discover/report) fraud unless the buyer is leaning on them to do so. An unsophisticated firm looking primarily at reach and impressions as key performance indicators sees those KPIs inflated by fraudulent impressions, so even the purchaser isn't really incentivized to discover fraud in those instances. And if you suddenly decide you do care about fraud after not caring about it previously, you're faced with the prospect of having to report to your stakeholders that the metrics you've been giving them have been bullshit for years.

If you're looking at conversions as your key performance indicator then you do care about fraud and will lean on your ad buyer to discover and reduce it, or else just avoid open web display inventory entirely and buy exclusively on platform properties like LinkedIn or Instagram.

There have been various efforts to reduce fraud for a long, long time. Obviously nothing so far has really comprehensively worked. But there are some people in the industry who do make their money and earn their reputations as fraud-fighters, and they continue pushing new solutions. If the fraud problem for open web display inventory could actually be solved it would probably result in a bonanza for anyone positioned to capitalize on a fraud-free supply chain, so some people are incentivized to chase the dream (or at least convince others that they're earnestly chasing the dream).

While I'm still waiting for the right moment in my own life to do my own Recurse Center retreat, I have had the pleasure of hiring and working with several RC alums over the years and in every instance have found them to be high-quality developers and interesting people to boot. Big congrats to RC for 15 years!

I recently tried to sign up for the Microsoft Partner Network and the entire experience was grueling and excruciating from start to finish. I wasted two days on it and never managed to sign up.

Before Claude Code, out of the alternatives you mentioned, I'd only tried Windsurf. I find Claude Code to be roughly comparable to Windsurf in terms of capabilities, but I can use it from my terminal. I'm much more at home in a terminal than I am in VSCode (which Windsurf is a fork of).

Doing row-by-row inserts into DuckDB is really slow. Accumulating rows in an in-memory data structure and periodically batching them into something like an in-memory Arrow table, and then reading the Arrow table into DuckDB, is fast and has been tenable for my own use cases.

I competed against IBM for many contracts over many years as an engineer at a small startup. The experience left me with zero respect for IBM. On occasions when a customer did choose IBM over us for a project, I would genuinely feel bad for the customer.

IBM is all smoke and mirrors; sales, marketing, and predatory business practices once they get their claws in you. Very little deep technical competence at that company.

I think a big part of an engineering manager’s job is protecting their team’s focus, and advocating for their team to the rest of the organization. Programmers should be insulated by an engineering manager from the vicissitudes and stresses of corporate politics, and their work should be protected from management-level priority thrashing.

That the Russian government is characterizing these personnel as “peacekeepers” would be a fair statement, but to regurgitate that characterization in reporting on the matter seems rather less credible.