Post

You should know that a big part of 18F's work was to make sure multi-million to multi-*hundreds*-of-millions dollar contracts at fed *and* state level didn't go to shitty enterprise IT consultancies that *repeatedly* delivered tech that didn't work, was late, or didn't even do what it needed to

Every one of those consultancies (e.g. an easy target is Deloitte -- you can look up what they've been involved in) would blame government for being a bad customer, and 18F were *also* involved in massively improving government's ability to run + manage these services

I brought 18F in on its first engagement with a state. They helped us turn around a giant half-billion-dollar procurement in a waterfall RFP into something that was broken up into smaller chunks in < 4 weeks. They stood up user research. They helped assess vendors on what they could *actually* do.

They help states check whether consultancies and bidders can do what they say they do -- because many governments lack the experience + capability to check those claims anymore. When people say they were the tip of the spear, they really were.

18F is where I met Robin Carnahan -- she was in charge of figuring out how a federal cost recovery org could even *work* with states. Robin would later go on to *run* the GSA in part because she know how important tech is to deliver government that works for the people.

Every single time I'd work with a government, I'd be able to point to 18F's derisking guide to help government leaders understand that a) how they were doing tech wasn't working (which they knew even if they wouldn't admit it) but most importantly b) that there's another way and they could get help.

You can't imagine how difficult it is to persuade a government to fundamentally change how it develops/acquires/manages technology at any scale without the reassurance that the federal government will support it. States needed 18F to even have the thought that feds would have their back.

I will be the first to admit that I've had problems with 18F and wish it did things differently or better at times, but I can say without reservation that getting rid of them is one of the strongest signs that all this bullshit is politically motivated, in case you still needed to be persuaded.

If a criticism of 18F is that it didn't go faster (some bullshit on LinkedIn that they "failed to audit the Pentagon 7 times"), a) they were no more than 100 people, b) they actually realized the way through this is to bring all of government along, and not through diktat.

There are people out there building tech for government that works, that is accessible, that isn't a ripoff, and many of those companies exist *because* 18F helped government become a more sophisticated buyer. Not even a buyer but a partner that takes better ownership + responsibility.

People with experience 18F would fan out to state and local governments too, where arguably better tech in government could affect even more people, just because of how government works in this country and states are closer to delivery.

And I *know* that one of the reasons why the Musk + Trump admin killed 18F is because their slack had a bot that reminded you if you used the word "guys" that perhaps you could use a more inclusive term and *sure* some of you might roll your eyes at that but *it wasn't doing any fucking harm*

And as I understand it 18F achieved its successes *despite* the fact that it was cost recovery. Agencies + departments would have to pay for 18F and they'd only do that *if they knew it would be better value for money*.

You wanted government that works for the people? For the first time, 18F was supporting deep efforts throughout gov to build tech *by talking to the people who use it* and learning from them. And not just build tech, but deliver policy! Which gov is terrible at doing! But now Musk knows better.

You know the kind of stuff 18F supported across government? It supported figuring out how *and whether* government can use AI without fucking you over. Whether or where government should use AI in the first place.