@mheadd@mastodon.social
@mheadd@mastodon.social avatar

mheadd

@mheadd@mastodon.social

Civic technology, open data, public service.

#CivicTech #government #technology #AI #ArtificalIntelligence #ServiceDelivery #PublicService #GovTech

This profile is from a federated server and may be incomplete. Browse more on the original instance.

mheadd, to random
@mheadd@mastodon.social avatar

"LLMs are useful tools for thought. They are terrible tools for delegating decision making to. That's currently my red line for using them: any time someone outsources actual decision making authority to an opaque random number generator is a recipe for disaster."

https://arstechnica.com/information-technology/2024/06/zoom-ceo-envisions-ai-deepfakes-attending-meetings-in-your-place/

mheadd, to ai
@mheadd@mastodon.social avatar

This is a fundamental mistake that people make when trying to assess whether LLMs are an appropriate tool to use in optimizing a process, function, or service:

"LLMs are not search engines looking up facts; they are pattern-spotting engines that guess the next best option in a sequence."

This terrific article is a great explainer on how they work and their limitations.

https://ig.ft.com/generative-ai/

#AI #ChatGPT #GenerativeAI

mheadd, to random
@mheadd@mastodon.social avatar

Watching this new site by @skinnylatte with great interest. If you care about how AI is used by government agencies, you should too. https://publicsectorai.tech/

mheadd, to random
@mheadd@mastodon.social avatar

So incredibly proud and grateful to have had the opportunity to serve with the fine folks at 18F. I learned so much, but the thing that has stayed with me is that we still have a lot of work left to do. LFG!

https://18f.gsa.gov/2024/03/19/18f-at-ten/

mheadd, to random
@mheadd@mastodon.social avatar

Love this blog post about writing blog posts. You learn so much about what you think you know when you have to write it, and you know the audience will try to make use of the information you share. The same feelings got me started writing my first blog. https://jack-vanlightly.com/blog/2022/1/25/write-for-others-but-mostly-for-yourself

mheadd, to ai
@mheadd@mastodon.social avatar

Authored more than 20 years ago, more prescient words have not been written about the choices governments face today in adopting new AI and LLM tools.

"The choices we face in the present regarding the use of digital tools and the institutional arrangements in which they are embedded will influence the way governments work around the globe during the next century and beyond."

  • Building the Virtual State, Jane E. Fountain (2001)

mheadd,
@mheadd@mastodon.social avatar

The spread of Internet in the late 90's and the ubiquity of the web browser forever changed the way that governments work. The choices that governments made then (and still do today) about how these technologies are used informed the shape of that change - who it benefitted, and who it didn't. The same will be true for the choices governments make in adopting and using AI and LLMs.

I have a new blog post coming out next week that will dive into this in more detail. Can't wait to share it!

mheadd, to eagles
@mheadd@mastodon.social avatar

Happy "Cowboys Choke in the Playoffs Day" to all who celebrate.

mheadd, to random
@mheadd@mastodon.social avatar

The real heroes in government are the people working through complex, poorly documented processes and writing down what they are doing so that others may follow.

Blessed are the documenters, for they will be called children of God.

mheadd, to accessibility
@mheadd@mastodon.social avatar

Being compliant with Section 508 and other requirements is the minimum we should be doing to provide modern digital services to people that need them, and that currently face challenges in accessing them. Time to start thinking beyond just being compliant.

https://adhocteam.us/playbook-accessibility/

mheadd, to random
@mheadd@mastodon.social avatar

"Don't tell me you love developers. Show me your platform documentation and I'll tell you what you think of developers."

mheadd, to random
@mheadd@mastodon.social avatar

People doing digital transformation in government: do not assume - and I can't stress this enough - that existing processes & requirements in an agency that keep teams from shipping good products are "broken."

They are not broken. They were designed with intent, and they work as designed. Even if they don't work to support shipping good software, they work for somebody.

Find out who, and how they benefit from existing processes & requirements and you are on your way to making lasting change.

mheadd, to random
@mheadd@mastodon.social avatar

Saying "thank you" is both free and priceless.

mheadd, to random
@mheadd@mastodon.social avatar

New on the Ad Hoc blog - Plain language for the win: Improving customer experiences with clearer communication. https://adhoc.team/2023/11/08/plain-language-customer-experience/

elizayer, to random
@elizayer@mastodon.social avatar

OK I’m gonna say it -

I don't think “Failure as an opportunity for learning” is ever going to resonate in government. The stakes are too high for that to feel true for most civil servants.

Failure might also mean that some people don’t eat. Or get medicine. It might mean that a whole program to support these things gets cancelled. The language that’s useful for the private sector around failure just doesn't feel respectful of the sacred duty government folks often feel.

mheadd,
@mheadd@mastodon.social avatar

@elizayer maybe replacing "failure" with "iterating quickly"?

The thing is, though, that lots of governments are not set up (yet) or mature enough in their delivery practices to push out changes rapidly. Really underscores the need to develop and maintain these practices within agencies.

baldur, to random
@baldur@toot.cafe avatar

“The Techno-Optimist Manifesto | Andreessen Horowitz”

This manifesto is one of the more horrific documents I’ve read lately, largely because of how normalised its rhetoric has become

It’s fascism. Like falangism with the catholicism swapped out for the prosperity doctrine https://a16z.com/the-techno-optimist-manifesto/

mheadd,
@mheadd@mastodon.social avatar

@baldur For an historical perspective on this market fundamentalist drivel, I highly recommend "The Big Myth," by Naomi Oreskes and Erik Conway. https://www.bloomsbury.com/us/big-myth-9781635573572/

mheadd, to random
@mheadd@mastodon.social avatar

Fixed this headline about government chatbots.

mheadd, to random
@mheadd@mastodon.social avatar

"These numbers underscore how much poverty is a policy choice." https://www.washingtonpost.com/dc-md-va/2023/09/12/us-poverty-rate-census-uninsured-2022/

mheadd, to random
@mheadd@mastodon.social avatar

"The cycle repeats itself: we invest our hard-earned time and sometimes money to build on top of friendly, seemingly benign platforms—only to see those platforms wriggle out from under us and morph into something entirely different (and for our purposes, much worse)." https://www.spicyweb.dev/farewell-jamstack/

mheadd, to random
@mheadd@mastodon.social avatar

I have been remiss in calling out my friend and two-time former co-worker @tbaxter for all his work on the recently released Platform Smells Playbook. Wouldn't have happened without him. https://adhocteam.us/platform-smells/

mheadd, to random
@mheadd@mastodon.social avatar

This is super important for anyone working on government technology modernization to read. A comprehensive overview of mainframes. Of particular importance - the part on "why mainframes survive." https://arstechnica.com/information-technology/2023/07/the-ibm-mainframe-how-it-runs-and-why-it-survives/

mheadd, to random
@mheadd@mastodon.social avatar
mheadd,
@mheadd@mastodon.social avatar

@benb True. Sometimes...

mheadd, to random
@mheadd@mastodon.social avatar

The ethics guidance I got when I worked as a fed - that I should decline any gift, even if technically legal and regardless of how small, if it could present even the slightest appearance of a conflict of interest - still ringing in my ears.

mheadd, to random
@mheadd@mastodon.social avatar

"The current 'existential threat' framing [of AI technologies] is effective because it fits on a rolling news ticker, diverts attention from the harms being created right now by data-driven and automated technologies and it confers huge and unknowable potential power on those involved in creating those technologies." https://rachelcoldicutt.medium.com/on-understanding-power-and-technology-1345dc57a1a

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • provamag3
  • mdbf
  • ngwrru68w68
  • modclub
  • magazineikmin
  • thenastyranch
  • rosin
  • khanakhh
  • InstantRegret
  • Youngstown
  • slotface
  • Durango
  • kavyap
  • DreamBathrooms
  • JUstTest
  • GTA5RPClips
  • ethstaker
  • normalnudes
  • tester
  • osvaldo12
  • everett
  • cubers
  • tacticalgear
  • anitta
  • megavids
  • Leos
  • cisconetworking
  • lostlight
  • All magazines
    •