iiinova — where your digital self keeps evolving

Marty Santalucia @msantalucia.bsky.social · 11/08/2026

To everyone worried about Anthropic watermarking text because it will "expose" you as an AI user: don't worry, we already knew.

000

Marty Santalucia @msantalucia.bsky.social · 01/08/2026

Who else has some crazy Saturday night plans lined up?

000

Marty Santalucia @msantalucia.bsky.social · 30/07/2026

1/ In their hesitancy to embrace AI, Democratic political institutions are abandoning the next major information battlefield. And our reasoning is increasingly disconnected from electoral realities.

110

Marty Santalucia @msantalucia.bsky.social · 11/11/2025

Now do comments on listings, you cowards

030

Marty Santalucia @msantalucia.bsky.social · 31/10/2025

The three categories of Trump admin activity.

030

Marty Santalucia @msantalucia.bsky.social · 17/10/2025

Nah, they've already made up their minds...

030

Marty Santalucia @msantalucia.bsky.social · 05/10/2025

I check Twitter occasionally for professional reasons - ignoring an online space doesn't mean it isn't consequential. The difference from a few years ago is stark. My primary news source for years is now full of wild conspiracy theories, overly racist/hateful content, and internet meme garbage.

010

Marty Santalucia @msantalucia.bsky.social · 04/10/2025

I'm horrified by these images. Thank you, Alderperson Fuentes, for standing up for Chicago. My team and I worked with Jessie in 2023 and saw how she turned struggle into motivation for change every day. Today she proved again that the 26th Ward is in good hands.

030

Marty Santalucia @msantalucia.bsky.social · 03/10/2025

Just finished @swordsjew.bsky.social's book Wild Faith this morning, then saw this. We've got some real, deep-rooted, problems in this country that I don't think we're in a position to deal with.

010

Reposted by Marty Santalucia

Marty Santalucia @msantalucia.bsky.social · 14/09/2025

I just wouldn't want my, or someone close to me's death used like this. The political world will take everything if you let it. Here's proof they'll take the memory of you too. But perhaps that's a fitting end for Kirk - having his family's grief plundered for the best performing email that month

021

Reposted by Marty Santalucia

Governor JB Pritzker @govpritzker.illinois.gov · 18/09/2025

A free and democratic society cannot silence comedians because the President doesn’t like what they say. This is an attack on free speech and cannot be allowed to stand. All elected officials need to speak up and push back on this undemocratic act.

17376057817516

Reposted by Marty Santalucia

Joseph Menn @joemenn.bsky.social · 16/09/2025

New research shows #DeepSeek suggests less-secure code when it is asked to help groups out of favor with the Chinese government. With its open-source model being adopted widely, this soft influence and hackability could spread. Gift link with email address etc. wapo.st/46jEZrb

wapo.st

AI firm DeepSeek writes less-secure code for groups China disfavors

Research by a U.S. security firm points to the country’s leading player in AI providing higher-quality results for some purposes than others.

68847

Marty Santalucia @msantalucia.bsky.social · 14/09/2025

021

Reposted by Marty Santalucia

Jeff Sharlet @jeffsharlet.bsky.social · 10/09/2025

If I have any credibility with you from 20 years of reporting on rightwing & fascist movements, please listen when I beg you not to celebrate Charlie Kirk getting shot. Leave aside morality: this isn’t a match in dry grass, it’s a torch. We do not want what this could ignite & we would not win.

45574971845

Marty Santalucia @msantalucia.bsky.social · 03/09/2025

Let's talk about training data in AI...

020

Reposted by Marty Santalucia

Ari Cohn @aricohn.com · 28/08/2025

I dunno, when you start blaming OpenAI because unwell people did unwell people things, filing massive lawsuits, and waging ap press campaign, this seems the entirely predictable course correction. You wanted guardrails to ensure mentally ill people can't encounter harmful output, now you got it.

4114230

Marty Santalucia @msantalucia.bsky.social · 24/08/2025

Installing some games on the kids ipad. Suck it Roblox

110

Marty Santalucia @msantalucia.bsky.social · 24/08/2025

Match the moment for fucks sake. The military is being marched into US cities and this sounds like a bureaucratic disagreement. Those of us on the ground will never be able to make the case that this is a moment of existential national crisis if leadership refuses to communicate effectively.

100

Marty Santalucia @msantalucia.bsky.social · 23/08/2025

We are built to see life in shadows.

000

Marty Santalucia @msantalucia.bsky.social · 21/08/2025

Do I like this? No. Is this the right move? 110%.

010

Marty Santalucia @msantalucia.bsky.social · 21/08/2025

I just prepped 200 donor prospect calls for a client, including finding phone numbers and writing 1 paragraph bios for each. It took me 45 minutes. While I was eating lunch with my kids. If you aren't seeing returns on AI usage, you're using the tool wrong.

000

Marty Santalucia @msantalucia.bsky.social · 19/08/2025

While the White House keeps hammering on less regulation and faster AI development, voters actually want STRONGER guardrails. Who is the White House plan actually for? People want to know the information they're getting is authentic and trustworthy. www.govtech.com/artificial-i...

govtech.com

Report Finds Bipartisan Agreement on Government AI Regulation

A survey conducted by the Program for Public Consultation at the University of Maryland found that both Republicans and Democrats support government AI regulation. The new AI Action Plan could limit t...

131

Marty Santalucia @msantalucia.bsky.social · 19/08/2025

This week's #TheImpact newsletter is live! Beat the curve on AI policy and news. This week was all about (gas) power and new AI pushes from the federal government. Plus, a new way to get caught up from MFStrategies! Read The Impact here: mfstrategies.beehiiv.com/p/are-300-ta...

mfstrategies.beehiiv.com

Are 300% Tariffs are coming? Plus Open‐Weight AI, Gas Power, and a $1 On‐Ramp

AI plan speeds open‐model buildouts; West Texas pitched as data hub; public backs guardrails; leadership shifts; $1 Claude pilots; 300% chip tariffs loom—reshaping costs, power, and timelines.

000

Marty Santalucia @msantalucia.bsky.social · 14/08/2025

Me going to Grok when ChatGPT tells me it won't help me build my web scraper that evades detection better.

000

Marty Santalucia @msantalucia.bsky.social · 14/08/2025

What are the chances these healthcare plays are all open appeals to the Trump administration for contracts that will consume the remaining Medicare dollars?

010

Marty Santalucia @msantalucia.bsky.social · 13/08/2025

I'm very familiar with this feeling

000

Marty Santalucia @msantalucia.bsky.social · 12/08/2025

This is straight Yarvin language. What it doesn't say is that, to Yarvin, CEOs are kings.

57227

Marty Santalucia @msantalucia.bsky.social · 12/08/2025

Minutes? Minutes?! Don't lie to the children. A three minute song required committing an entire evening to the project.

010

Marty Santalucia @msantalucia.bsky.social · 11/08/2025

If only the government had some power to fairly and consistently collect money from companies without extortion. If only!

nytimes.com

U.S. Government to Take Cut of Nvidia and AMD A.I. Chip Sales to China

010

Reposted by Marty Santalucia

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 10/08/2025

Dang it @gracekind.net beat me to it bsky.app/profile/grac...

0111

Reposted by Marty Santalucia

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 08/08/2025

Every time you share the “3 Bs in blueberry thing” and no one explains its relation to tokenizers, you help everyone get just a tiny bit less informed

2218320

Marty Santalucia @msantalucia.bsky.social · 08/08/2025

Gonna be honest that I kinda hate OpenAI's direct targeting of healthcare. Startup people are always trying to get a piece of that pie because "Total addressable market". So they're running into the fire with a tool they openly admit makes shit up *less* than previous models (Notably not "never").

010

Reposted by Marty Santalucia

Mike Masnick @masnick.com · 06/08/2025

I feel like every single day I need to post a reminder: if you ask an LLM about itself it DOES NOT KNOW ANYTHING ABOUT ITS INNER WORKINGS and WILL NOT TELL YOU ANYTHING ACTUALLY ABOUT ITSELF. It just (as it always does) will make up some plausible sounding answer.

834493958

Marty Santalucia @msantalucia.bsky.social · 28/07/2025

The idea that all government activities must turn a profit is not just as philosophical difference of opinions. It is fundamentally (and likely deliberately) misunderstanding what the government does and how it creates value.

010

Marty Santalucia @msantalucia.bsky.social · 27/07/2025

Sad news. Tom Lehrer and Monty Python are responsible for at least 75% of my worldview.

nytimes.com

Tom Lehrer, Musical Satirist With a Dark Streak, Dies at 97

000

Marty Santalucia @msantalucia.bsky.social · 14/07/2025

Oh ok. We're giving MechaHitler access to the bombs... But, seriously, the lack of transparency on *how* this system will be used is incredibly concerning. Will an LLM be making life or death decisions? We don't know.

000

Marty Santalucia @msantalucia.bsky.social · 13/07/2025

And if you do "pick yourself" they will find their own candidate against you. They also "pick" the staff and threaten candidates by withholding millions of dollars. If you still beat them (I have), they are incredibly disrespectful and explicit about how unhappy they are with the result.

100

Marty Santalucia @msantalucia.bsky.social · 11/07/2025

Without proper regulation, a primary benefit of rolling out AI in public spaces will be to avoid accountability for unpopular decisions. www.politico.com/news/magazin...

I would start by saying companies that make large language models need to be held responsible in some way for the things those systems say, which includes defamation, hate speech, etc. And right now, legally, they’re not. It’s not really clear that the companies are responsible for what those systems do. Also, it’s not clear that it’s responsible if those systems plagiarize — that piece of the law is very much open right now.

010

Marty Santalucia @msantalucia.bsky.social · 09/07/2025

Let's imagine that Grok is tasked with making decisions for government agencies. We dump in data and then the black box decides how to distribute, let's say, healthcare money. It won't be verbose, just outputting "approved" or "declined". But could this be the wiring behind the scenes?

020

Marty Santalucia @msantalucia.bsky.social · 01/07/2025

Cardinal rule of small children: there is never just one puddle.

000

Marty Santalucia @msantalucia.bsky.social · 30/06/2025

Even by current "standards" this is dizzying. What the fuck are we doing?

010