In this episode, I talk with Ramon Perez, Executive Director of the Digital Democracy Project, a nonpartisan nonprofit using secure mobile voting technology to give citizens a real-time voice in legislation. Ramon explains how the platform lets verified, registered voters weigh in on bills being debated in Congress and their state house—and then scores legislators on how closely their votes match what their districts wanted. We dig into how AI, including a RAG-powered chatbot called VoteBot, helps everyday citizens parse thousands of pages of complex legislative text. We also discuss digital security, participatory budgeting, and Ramon’s ambitious goal of expanding the platform to all 50 state legislatures by 2027.
Resources
Digital Democracy Project website | Votes (VOATZ) app
Guest Bio
Ramon Perez Ramon Perez is the Executive Director of the Digital Democracy Project, a nonpartisan nonprofit organization that uses secure mobile voting technology and AI to enable citizens to vote on legislation and hold their elected officials accountable. He has worked in AI and machine learning for more than 12 years and previously interned at the Congressional Budget Office as a Georgetown Economics graduate student.
Different Ways to Listen to the Show!
New Ways to Support the Show!
With more than 300 guests and 12 seasons of episodes, the PolicyViz Podcast is one of the longest-running data visualization podcasts around. You can support the show by downloading and listening, following the work of my guests, and sharing the show with your networks. If you’re interested in financially supporting the show, you can sign up for my Patreon platform, make a one-time payment via PayPal, or shop one of the show’s sponsors.
Transcript
WEBVTT
1
00:00:04.630 –> 00:00:07.979
Jon Schwabish: Hello, Ramon, welcome to the show. Good to meet ya.
2
00:00:08.160 –> 00:00:11.089
Ramon Perez: Yeah, good to have… good to meet you as well. Thank you, John, I appreciate being on.
3
00:00:11.090 –> 00:00:29.360
Jon Schwabish: Yeah, excited to talk to you about your work. We’ve already been chatting a lot before we get out about AI, so we’ve got a lot to talk about. But why don’t we start with introductions? Maybe you can talk a little bit about yourself and your background and, you know, get us into the Digital Democracy Project.
4
00:00:30.210 –> 00:00:37.169
Ramon Perez: Yeah, sure, yeah, so I’m Ramon Perez, I’m the Executive Director of the Digital Democracy Project, which is a nonpartisan non-profit organization.
5
00:00:37.300 –> 00:00:47.059
Ramon Perez: We use mobile voting software, which has been used by, for military voters overseas to cast a secure absentee ballot on their phones.
6
00:00:47.210 –> 00:01:06.960
Ramon Perez: But we’re using it for a different purpose. We’re not asking people to vote in an election, we’re asking people to tell their legislators how they want them to vote on every bill that’s being debated in Congress or in their state house. And then we provide those results in real time, so that legislators know what their districts want them to do.
7
00:01:07.140 –> 00:01:16.579
Ramon Perez: And at the end of the legislative session, we compare what did the voters want in the app to what did the legislators do on the floor in the legislative chambers.
8
00:01:16.900 –> 00:01:34.539
Ramon Perez: And give every one of them a score, like a baseball card, that shows how often do they match their district. So, the concept is very simple. It’s vote, tally, score. You vote on the same bills as your legislators. We tally up the aggregate results and show them on the website as they’re coming in.
9
00:01:34.890 –> 00:01:38.050
Ramon Perez: And then we score the legislators against what do the districts want.
10
00:01:38.240 –> 00:01:47.549
Ramon Perez: The key thing that makes it possible, as I mentioned, is that mobile voting software. We partnered with a company called Votes, which is V-O-A-T-Z, based out of Boston.
11
00:01:47.690 –> 00:01:52.779
Ramon Perez: And they’ve been building this election software now for about 10 years.
12
00:01:52.800 –> 00:02:09.150
Ramon Perez: And, have had about 5 million votes cast in about 150 different elections, and these are binding elections, and several states now are doing this. They’ve all… a lot of the states who have piloted this, and red and blue states, mind you, so I think they started in West Virginia.
13
00:02:09.300 –> 00:02:14.150
Ramon Perez: but have worked in Utah, Arizona, Oregon, Michigan, Massachusetts.
14
00:02:14.430 –> 00:02:25.380
Ramon Perez: I think some in Pennsylvania, I can’t remember, a variety of states, and… oh, Denver did want a municipal, and now they’re doing…
15
00:02:26.160 –> 00:02:31.579
Ramon Perez: Now, I think the government of Mexico, in the last national election, allowed
16
00:02:31.710 –> 00:02:51.419
Ramon Perez: all Mexican citizens outside the country to just vote from their phone using the Votes app. So, in Canada, in Ontario, the Ontario provincial government said that for municipal elections, anybody can vote from their phone. So now a bunch of cities have been allowing people to vote from their phone.
17
00:02:51.530 –> 00:03:02.069
Ramon Perez: In Ontario. So, I think the technology itself is exciting, and as I mentioned, we’re not a… we’re not a supervisor of elections, so we’re not running a binding election, but we…
18
00:03:02.210 –> 00:03:04.130
Ramon Perez: figured,
19
00:03:04.300 –> 00:03:10.710
Ramon Perez: why couldn’t we use this technology to build something like the Netflix of government, where you have government on demand, instead of having people
20
00:03:10.880 –> 00:03:25.079
Ramon Perez: go to an election once every couple years, and then sit on their hands the rest of the time, like a bump on the log, just shaking our fists at the TV, we could have a say in all of the most consequential legislation that affects our lives.
21
00:03:25.080 –> 00:03:34.100
Ramon Perez: So, the Iran War Powers Resolution, for example, was one we carried recently. Voting on whether or not to go to war, to me, seems like the most
22
00:03:34.150 –> 00:03:53.650
Ramon Perez: consequential reason to ask voters what they want to do. And also, I think the key is to hold legislators to account for the results. I think that’s the main thing that we’re doing that is different from just sending your congressman an email or calling the switchboard, is that there’s transparency in saying.
23
00:03:53.670 –> 00:03:54.600
Ramon Perez: you know.
24
00:03:54.830 –> 00:04:04.520
Ramon Perez: We verified voters against the voter file, we verified them with a photo ID that matches their face in a selfie scan. We know they’re real human beings, we know they’re registered in this district.
25
00:04:04.710 –> 00:04:19.019
Ramon Perez: Let’s give them the right to have a say, and let’s then compare apples to apples to say voters wanted X, and the legislator voted Y, and… and hold them accountable at come election time.
26
00:04:19.630 –> 00:04:20.220
Jon Schwabish: Right.
27
00:04:20.350 –> 00:04:30.389
Jon Schwabish: So, can you explain a little bit, what happens on the… with varying sample sizes? So, you’ve got one person’s district.
28
00:04:30.700 –> 00:04:36.329
Jon Schwabish: 80% of people register on the app, and another person’s district, you know, 3 people register on the app.
29
00:04:36.500 –> 00:04:49.190
Jon Schwabish: How does the, the, I guess the… the view that the public would see, or the legislator, more importantly, the legislator would see, take into account the different potential sample sizes?
30
00:04:49.870 –> 00:04:59.029
Ramon Perez: Yeah, well, we take, we are not a polling agency, so we’re not trying to find and selectively sample a likely voter.
31
00:04:59.180 –> 00:05:17.310
Ramon Perez: group, if you will. Our goal is to provide a platform so everybody can participate as they choose, and we do just report out the numbers. So if there are very few people in a given district who have voted, that is visible on the scorecard so that you get some sense of, okay, what’s the…
32
00:05:17.480 –> 00:05:35.749
Ramon Perez: what’s the magnitude of user input that we’ve gotten here? There are lots of ways that I… I mean, I think you’re the policy, you know, visualization expert. I think there are a lot of ways that you could probably help us make better visuals that can take advantage of this data, but our goal is radical transparency. We want to provide all the data.
33
00:05:35.750 –> 00:05:38.699
Ramon Perez: We want to let people make their own minds up, but it’s not…
34
00:05:38.720 –> 00:05:55.429
Ramon Perez: it’s not… it’s not as if we are saying, we’re going to sample from these people’s opinions, and we’re going to elevate their opinions and suppress these opinions in order to get something that looks like what we call likely voters. We’re not doing that. The other thing, too, is that we’re asking people about bills. We’re not asking
35
00:05:55.430 –> 00:05:59.030
Ramon Perez: About a horse race, you know, would you vote for, you know.
36
00:05:59.050 –> 00:06:02.790
Ramon Perez: Harris or Trump type of thing that a typical polling company does.
37
00:06:02.940 –> 00:06:11.049
Ramon Perez: We’re saying, do you, if you, you know, having read this, Homeland Security funding bill.
38
00:06:11.760 –> 00:06:26.930
Ramon Perez: do you think that this is something that should be passed, yes or no? And some people are going to be very interested in that Homeland Security funding bill, and other people will not care one iota. Right. So, you’re gonna find that
39
00:06:27.280 –> 00:06:37.000
Ramon Perez: that, you know, John, you might be… you might care quite a lot about healthcare policy, and I might care quite a lot about taxes, but…
40
00:06:37.130 –> 00:06:54.860
Ramon Perez: But you may only interact on a healthcare bill and ignore all the rest of them, and I may only care about a tax bill and ignore all the rest. You have some people who are super users who are going to want to vote on every single thing. We have those folks. Every bill we load, they will go in and they’ll read about it and they’ll vote on it, which is awesome.
41
00:06:54.970 –> 00:07:09.980
Ramon Perez: So we… our goal is to provide a platform and scale the platform, rather than trying to find the perfect set of people that can represent opinions, like a typical polling, company would do.
42
00:07:10.980 –> 00:07:15.219
Jon Schwabish: So, let me ask on the piece you just mentioned about… about the bill.
43
00:07:15.370 –> 00:07:32.830
Jon Schwabish: Do you, or have you thought about trying to generate some sort of summary? I mean, these bills, I mean, the budget bill, for example, is tens of thousands of pages. Yeah. Have… does the tool, or have you thought about trying to provide people with a summary of what’s in there?
44
00:07:33.320 –> 00:07:37.470
Ramon Perez: Yeah, no, it’s great. I’m glad you asked that, because that’s another key,
45
00:07:38.320 –> 00:07:54.750
Ramon Perez: technological development which makes this project possible, which is AI. Because we started this… we started the project in… we’ve started going into beta testing in 2022, and did our first live legislative session. It was a Florida legislative session in 2023.
46
00:07:54.820 –> 00:08:01.939
Ramon Perez: And in that time, what I did was I hired a Democrat and a Republican who were legislative policy analysts
47
00:08:02.130 –> 00:08:04.919
Ramon Perez: who had worked on Capitol Hill in Tallahassee.
48
00:08:05.430 –> 00:08:07.959
Ramon Perez: Then they would just read the bills and…
49
00:08:08.100 –> 00:08:12.489
Ramon Perez: Work together on a summary that we could agree would be nonpartisan but factually accurate.
50
00:08:12.610 –> 00:08:13.730
Ramon Perez: Yeah.
51
00:08:13.960 –> 00:08:23.590
Ramon Perez: That was very time-consuming. You’re talking thousands of pages, so we couldn’t carry that many bills, we only could carry, really, the high… the marquee legislation.
52
00:08:23.770 –> 00:08:29.910
Ramon Perez: And… Later on that day, year, I want to say it was.
53
00:08:30.400 –> 00:08:49.990
Ramon Perez: ChatGPT, or not just ChatGPT, but the ability to upload programmatically a document to the OpenAI API and get back a summary was groundbreaking. I mean, that happened that fall, I want to say. We introduced that very quickly, and we’ve been… my day job, I work in AI and machine learning.
54
00:08:50.200 –> 00:08:52.540
Ramon Perez: And have done for the last 12 years, so I am…
55
00:08:52.550 –> 00:09:03.319
Ramon Perez: A big believer in using these tools everywhere within the stack, but one of the main important things for us was summarizing complex legalese in a easily digestible.
56
00:09:03.320 –> 00:09:12.659
Ramon Perez: human-readable form, right? Because bills are not written for us. They are written by lawyers or other lawyers who are going to defend it in front of some judge someday.
57
00:09:12.660 –> 00:09:16.490
Ramon Perez: And so, they have to be incredibly pedantic in the way that they’re written.
58
00:09:16.620 –> 00:09:27.329
Ramon Perez: So, that makes them very, very difficult to interpret, oftentimes. And small changes in language can have a big impact on the law. So…
59
00:09:27.510 –> 00:09:30.920
Ramon Perez: AI has given us the incredible ability
60
00:09:30.960 –> 00:09:50.560
Ramon Perez: to let people know what is in the discussion, what’s in the debate here, what’s being considered in their names by their legislators. So if you… if you look on our website now, on digitaldemocracyproject.org, if you’re looking at any given bill, you’ll see this little chat bot in the bottom right, which we call VoteBot.
61
00:09:50.580 –> 00:10:09.660
Ramon Perez: VoteBot has been fed all of the legislative text from thousands of bills into a vector database in Pinecone, and then it runs a RAG query against it, so we can talk about AI if you’re interested in that kind of stuff, but it can basically rapidly read
62
00:10:09.660 –> 00:10:13.800
Ramon Perez: The text in response to a user’s question, and…
63
00:10:13.900 –> 00:10:27.220
Ramon Perez: and provide some really detailed answers, because it’s got the full body of the text available to it, which is something you don’t actually typically get from just a regular ChatGBT or Perplexity or Claude search.
64
00:10:27.450 –> 00:10:41.110
Ramon Perez: Because they don’t… they don’t preprocess the entire bill. I mean, it’s a lot of… it’s a lot in there. And you have to maintain versions, because the bill gets amended, you have to update with amendments as those come through, and that can happen
65
00:10:41.350 –> 00:10:56.440
Ramon Perez: pretty much every day during a legislative session, then once the session’s over, that bill’s never getting touched again. So it’s dead, or it became law. So it’s a lot of documents to manage, and we do that as part of our work, because it’s not just
66
00:10:56.600 –> 00:10:58.290
Ramon Perez: For Digital Democracy Project.
67
00:10:58.570 –> 00:11:07.249
Ramon Perez: as you rightly point out, John, it’s not just about being able to have a say in what mobile voting gets you in a secure mechanism for reaching out to your elected official.
68
00:11:07.260 –> 00:11:16.269
Ramon Perez: It’s being able to understand, it’s being able to educate yourself on the policies, and to the level where you have enough comfort to have a say.
69
00:11:16.270 –> 00:11:27.909
Ramon Perez: Because a lot of times what people will say to me is, well, you know, we elect these politicians to speak on our behalf because I don’t know as much as they do, right? They’re supposed to go to Capitol Hill.
70
00:11:27.910 –> 00:11:36.340
Ramon Perez: And they have a team, and they understand all these bills, and they have a fine-grained level of detail and can make an educated decision.
71
00:11:36.510 –> 00:11:44.210
Ramon Perez: And usually what I’ll say is, like, you haven’t spent much time with these guys. I think what you find is.
72
00:11:44.270 –> 00:12:02.359
Ramon Perez: their budgets have actually been cut so much over the last 30 years that oftentimes they’re sharing staff, even… especially at a state house level. And I’ve spent a lot of time in Tallahassee, watching legislators from both sides of the aisle go through the motions, some juniors, some seniors, some House leadership, etc.
73
00:12:02.990 –> 00:12:13.039
Ramon Perez: Same in DC. You know, what you find is a junior, congressman in Washington is expected to spend 70-80% of their time dialing for dollars.
74
00:12:14.080 –> 00:12:21.380
Ramon Perez: If they show up for a vote, the vote is gonna be a party-line vote. They’re gonna vote the way the House leadership and the Whip’s office tells them to vote.
75
00:12:21.430 –> 00:12:27.840
Ramon Perez: Oftentimes, they rarely ever read the bills, because their job is to go raise money for the party. It is not…
76
00:12:27.880 –> 00:12:42.260
Ramon Perez: to understand policy at a deep level. And Washington is much worse in this sense because, at least in the state legislatures, they tend to follow regular order and still have time for debate and committee hearings and whatnot.
77
00:12:42.260 –> 00:12:42.770
Jon Schwabish: Yeah.
78
00:12:42.770 –> 00:12:59.189
Ramon Perez: in DC, you know, they’ll roll out major policy at the 11th hour and have a voterama until 2 o’clock in the morning. Nobody can be expected to read the one big, beautiful bill in the 12 hours that they’ve had to look at. I mean, it’s not realistic.
79
00:12:59.610 –> 00:13:03.070
Jon Schwabish: So do you… does… does,
80
00:13:03.210 –> 00:13:09.840
Jon Schwabish: Do you pull from… for bills that are modifying existing law?
81
00:13:10.010 –> 00:13:15.849
Jon Schwabish: Do you also pull in the existing law? So, for example, you can see bills that say, you know.
82
00:13:15.950 –> 00:13:27.780
Jon Schwabish: this is modifying line 17 of blah blah blah to say from group X to Group Y. So, you have all the current votes, does it also look back at the existing law?
83
00:13:28.380 –> 00:13:31.840
Ramon Perez: That is actually a very insightful question. So…
84
00:13:33.620 –> 00:13:43.160
Ramon Perez: I would say not directly. We don’t pull existing law into the vector database for VoteBot. What… the way a legislation is marked up.
85
00:13:43.310 –> 00:13:46.870
Ramon Perez: They’ll usually annotate when something changes existing law.
86
00:13:46.870 –> 00:13:47.280
Jon Schwabish: Yeah.
87
00:13:47.280 –> 00:13:50.459
Ramon Perez: But they… but they don’t necessarily say.
88
00:13:50.600 –> 00:13:56.129
Ramon Perez: Basically, they’ll say, you know, this highlighted section is a change to law XYZ, right?
89
00:13:56.130 –> 00:13:56.660
Jon Schwabish: Yeah.
90
00:13:56.660 –> 00:14:01.159
Ramon Perez: But it doesn’t… it’s not the entire law, it’s just saying this section changed, or this line changed, right?
91
00:14:01.230 –> 00:14:17.959
Ramon Perez: So, I do think that would be a really fascinating, additional feature for us to build, and I would be… I’ve always had in my head where it would be great if we could mark up exactly what changed and show that in a visual way so that people can say.
92
00:14:17.960 –> 00:14:23.289
Ramon Perez: This is what changed, and this is how it affects either existing law, or this is brand new law.
93
00:14:23.290 –> 00:14:26.449
Ramon Perez: We surface what Congress gives us.
94
00:14:26.450 –> 00:14:43.509
Ramon Perez: But I know that there are way better ways to visualize what Congress gives, other than they’re just doing, like, a green line through, or a red markup, or a, like, a strike-through, which is not, you know, easy to necessarily interpret. So I think that would be a… that’s actually a great idea for additional feature development.
95
00:14:43.510 –> 00:14:57.540
Jon Schwabish: Yeah. The last question I want to ask on this topic is, as someone who used to work at the Congressional Budget Office, do you bring in the budget numbers as well? And I’m going to layer on one other question here.
96
00:14:58.800 –> 00:15:16.869
Jon Schwabish: have you tried to look at what… what people who are using the app are looking at? So I’m combining these questions because I’m curious what you think people using the app are most interested in when they’re reading the bill, as opposed to… I’m sure there are many that are like.
97
00:15:17.650 –> 00:15:22.990
Jon Schwabish: I mean, just like… just like legislators, right? Like, I’m for this, I’m against this wholesale. I don’t need to read.
98
00:15:22.990 –> 00:15:25.230
Ramon Perez: Yeah. That, yeah, yeah, yeah.
99
00:15:25.230 –> 00:15:34.799
Jon Schwabish: Right? But for those who want to go a little bit deeper, like, what are they… where are they looking? What is the thing that they’re finding? And then, of course, the budget numbers as well. So, a lot of.
100
00:15:34.800 –> 00:15:44.899
Ramon Perez: That is… no, that’s great. For one, John, I didn’t realize you were at the CBO, because I was actually… I actually interned at the CBO in 2014 as a Georgetown Econ grad student.
101
00:15:44.900 –> 00:15:45.410
Jon Schwabish: Okay, so…
102
00:15:45.410 –> 00:15:49.929
Ramon Perez: Maybe our paths crossed at some point.
103
00:15:50.260 –> 00:16:03.420
Ramon Perez: I think Doug Elmendorf was the head of the CBO at the time. Yeah, oh, that’s very cool. So, yes, I have always wanted to do something with the budget, and really what I would like to do is participatory budgeting.
104
00:16:03.570 –> 00:16:07.510
Ramon Perez: Where… So, the budget documents, we have not…
105
00:16:07.760 –> 00:16:15.460
Ramon Perez: Generally, what we do is we say, okay, the budget, we will care… we’ll bring it in, and people are voting up or down. Do you agree?
106
00:16:15.500 –> 00:16:27.940
Ramon Perez: passing this legislation, where the legislation is the budget. Do you agree with that, yes or no, all right? But really what you want is to get to… the real power, I think, of the technology would be allowing a lot of levels of nuance.
107
00:16:27.940 –> 00:16:39.289
Ramon Perez: Where rather than just a straight up or down vote from people, if we could use participatory budgeting, where we could say, okay, given this bucket of money for DHS, for example.
108
00:16:39.290 –> 00:16:51.339
Ramon Perez: how would you prioritize it? Like, don’t… let’s not… there’s a thousand line items, so we can’t expect people to read every one and say, I agree with that line item, but more so, if you had $100 to spend on the Department of Homeland Security.
109
00:16:51.350 –> 00:16:58.130
Ramon Perez: what would be the ways that you would assign the money? Okay, you take, you know, 17 of those dollars, and you put it towards, like.
110
00:16:58.130 –> 00:17:13.549
Ramon Perez: border security, or maybe you take 5 of those dollars and you put it towards the FAA. Maybe you take, you know, one of those dollars and you put it towards customs. I don’t know, whatever it is. And that would be really interesting, because then, over a broad swath of people, you could actually then
111
00:17:13.599 –> 00:17:26.480
Ramon Perez: aggregate enough information to get a crowdsourced sort of steer on what people’s priorities are for spending. I think that would be really, really fascinating, and…
112
00:17:26.760 –> 00:17:32.270
Ramon Perez: And the mobile voting, the votes app that we’re using, it…
113
00:17:32.390 –> 00:17:51.899
Ramon Perez: we would have to work with votes to build out that part, like what I have in mind. They could do participatory budgeting… they have ranked choice voting and advisory voting built in, so you could do kind of a… you could kind of do a hacky way of doing participatory budgeting in the form of a ranked choice vote, where people just sort of rank their priorities.
114
00:17:51.900 –> 00:18:01.419
Ramon Perez: But it’s one thing to say, okay, my number one budget spend would be border security, but that’s different from assigning a dollar amount, right?
115
00:18:01.420 –> 00:18:01.890
Jon Schwabish: Yeah.
116
00:18:01.890 –> 00:18:03.380
Ramon Perez: You know what I mean? Like, you could say.
117
00:18:03.380 –> 00:18:03.770
Jon Schwabish: Yeah.
118
00:18:03.770 –> 00:18:23.640
Ramon Perez: I want the FAA to be funded, but the FAA is a small proportion of the overall spend, so even if it’s your number one priority, that’s different from assigning a dollar amount, so I think that I would love to do something on participatory budgeting, and do it in a framework where we still maintain all the security and the whole blockchain backend that we’re using, and all the.
119
00:18:24.150 –> 00:18:35.619
Ramon Perez: the homomorphic encryption and all that kind of stuff, because there are participatory budgeting tools that exist on the internet, but nothing that combines the security of mobile voting. So, yeah, but I would love that.
120
00:18:36.090 –> 00:18:39.350
Jon Schwabish: Yeah, that’s very cool. I mean, I could see, yeah, lots of…
121
00:18:39.670 –> 00:18:42.280
Jon Schwabish: Pass forward into lots of these other
122
00:18:42.760 –> 00:18:49.760
Jon Schwabish: other data sets and pieces of information to pull in, pull in together. Okay, so,
123
00:18:49.980 –> 00:18:52.060
Jon Schwabish: One of the…
124
00:18:52.340 –> 00:19:00.100
Jon Schwabish: I laugh a little bit, but it’s actually quite… not… not funny, but one of the things that we’ve obviously had to discuss in the last…
125
00:19:00.210 –> 00:19:21.780
Jon Schwabish: six years is, voter fraud, in, in quotation marks, and we know, I think everybody listening to this podcast knows all these, so I’m not going to rehash all that. We know about the Dominion voting, we know about general critiques of, of digital and mail-in voting now, which of course is part of the SAVE Act that will get rid of a lot of the mail-in voting, so…
126
00:19:22.460 –> 00:19:27.910
Jon Schwabish: Your… your project is not… Actually, elections, so we’re not there.
127
00:19:28.480 –> 00:19:34.810
Jon Schwabish: I’m just curious about… how you think about, or how DDP thinks about.
128
00:19:34.930 –> 00:19:38.920
Jon Schwabish: Digital security, voting, elections, sort of just generally, like.
129
00:19:38.920 –> 00:19:39.640
Ramon Perez: Yeah, yeah.
130
00:19:39.830 –> 00:19:43.859
Jon Schwabish: And does the project try to maybe…
131
00:19:44.360 –> 00:19:53.300
Jon Schwabish: Like, an underlying effect of this to kind of ease people’s concerns, that we can use modern technology to do what we’ve been doing for 200-some-odd years in this country?
132
00:19:53.720 –> 00:19:56.050
Ramon Perez: Yeah, yeah, I think, I think…
133
00:19:56.930 –> 00:20:01.789
Ramon Perez: you ask some important questions to this, right? It’s because
134
00:20:01.860 –> 00:20:11.410
Ramon Perez: we could do Digital Democracy Project just in a regular web app. Like, we could, you know… and there are other tools out there that allow people to…
135
00:20:11.410 –> 00:20:25.099
Ramon Perez: have a say on legislation, and then, you know, you click a button, do you support this bill, and then it sends an email to your congressman. Like, those tools exist, right? Why do we spend so much money and time and energy trying to integrate this much more complex
136
00:20:25.170 –> 00:20:30.010
Ramon Perez: stack. And it’s because my feeling has always been that
137
00:20:30.450 –> 00:20:35.170
Ramon Perez: The goal is to tell legislators how to vote
138
00:20:35.680 –> 00:20:49.379
Ramon Perez: and provide them that information before they walk into the chambers and actually cast a vote in Congress. And the only way they’re going to listen to that is that they have high confidence that these are real human beings.
139
00:20:49.380 –> 00:21:00.570
Ramon Perez: they are registered to vote, and they live in my district, so they’re U.S. citizens, we can match their… you know, we’ve done all the front-end verification of IDs and documents, and…
140
00:21:00.990 –> 00:21:16.710
Ramon Perez: So that’s why I have always… I have never personally understood this concern about voter ID. In fact, from the latest polling I’ve seen, even a majority of Democrats and something like 80% of Americans overall believe that you should have an ID when it’s time to vote.
141
00:21:16.860 –> 00:21:25.279
Ramon Perez: Because you do need to prove who you are, I mean, you know, so… and for our system, it is… we… we simply…
142
00:21:25.340 –> 00:21:38.949
Ramon Perez: could not go without ID verification. Nobody would trust the results if anybody could create an account using grandma’s, you know, name and zip code and date of birth from the voter file. So…
143
00:21:39.350 –> 00:21:45.120
Ramon Perez: We have to… we… we have to spend a lot of money to build a system
144
00:21:45.270 –> 00:21:57.889
Ramon Perez: or rely on a lot of the money that Votes has spent also to build a system that incorporates a zero-trust framework and homomorphic encryption and protects the secret ballot, which is another important thing.
145
00:21:58.060 –> 00:22:13.010
Ramon Perez: And still provides real-time aggregated vote results in a way that matters for what we’re doing. So, I think security’s the number one responsibility. I think… I’ll often tell the team here that we’re not trying to build
146
00:22:13.180 –> 00:22:15.820
Ramon Perez: A product. We’re trying to build trust.
147
00:22:15.900 –> 00:22:33.929
Ramon Perez: We’re a community project. This is a team of all non… of all unpaid volunteers. Developers, legislative analysts, organizers, you know, volunteer, people volunteering to table an event, or knock on doors, or make phone calls. We have to…
148
00:22:34.010 –> 00:22:44.520
Ramon Perez: we have to build a project that the American people and our elected representatives will trust. And that means they have to believe
149
00:22:44.690 –> 00:22:48.310
Ramon Perez: From soup to nuts, they have to believe that we’re not putting
150
00:22:48.620 –> 00:22:55.899
Ramon Perez: shareholders in the way of any decision we make. We’re a non-profit organization.
151
00:22:55.900 –> 00:23:12.400
Ramon Perez: We are… that way we know, you know, there’s nobody with their thumb on the scales trying to manipulate the results. We have partnered with an org… with a company that has been state RAMP certified in several states, and… and gone through tons of security audits.
152
00:23:12.480 –> 00:23:17.429
Ramon Perez: we can’t actually see how people vote, it’s a secret ballot, so I can’t say
153
00:23:17.430 –> 00:23:38.040
Ramon Perez: All I know is that, you know, John Schwabish verified with a photo ID and is registered to vote in, I don’t know, Virginia, right? But I can’t see that you voted in favor of the Iran War Powers Resolution, and I voted against it. We can’t see that because it’s a secret ballot, and that’s exactly the way we like it, right? We have to protect people’s
154
00:23:38.040 –> 00:23:39.550
Ramon Perez: Privacy while still
155
00:23:39.550 –> 00:23:48.800
Ramon Perez: verifying security. So, so many of the things that we’ve done, which feel like, handcuffs sometimes, when people come to us and say.
156
00:23:48.920 –> 00:24:07.949
Ramon Perez: well, why is it that you can’t give me some specialized report on how this demographic supports this legislation, or whatever? And I say, I’m sorry, like, that’s… that’s a policy choice by design, because we’re not here to make money, we’re not… we are here to build trust for,
157
00:24:08.160 –> 00:24:17.519
Ramon Perez: to try to build a better version of democracy that is more accountable and more transparent for the American people. So yeah, I do agree with you that
158
00:24:18.040 –> 00:24:23.939
Ramon Perez: We’ve… we’ve sort of lowered the heat a bit on mobile voting, because
159
00:24:24.300 –> 00:24:37.400
Ramon Perez: this isn’t a binding election, so we’re not the supervisor of elections office, so we’re providing advisory voting to legislators, but I happen to also believe that as people use the technology, become more comfortable with it.
160
00:24:37.400 –> 00:24:49.569
Ramon Perez: they’re gonna say, gosh, well, like, why… this is so easy. I can sit here on my couch with my cup of coffee, and I can read about some complicated policy, and I’ve got this AI bot that’s telling me all the details about
161
00:24:49.590 –> 00:24:53.800
Ramon Perez: whether or not the FAA gets funded or not, and what’s the, you know.
162
00:24:54.130 –> 00:25:08.409
Ramon Perez: why shouldn’t I be able to vote like that? Like, why do I have to go into… the poor voters of California this year are gonna walk into a polling booth… well, actually, they do absentee vote by mail, but they’re gonna have something like 30 ballot initiatives on their ballot.
163
00:25:09.720 –> 00:25:26.459
Ramon Perez: How great would that be if you had a little chat bot that had read those ballot initiatives for you and could help you understand the fine details of the implications of every one of those proposed constitutional amendments or citizen, you know, citizen-proposed amendments? Several states have
164
00:25:26.520 –> 00:25:40.350
Ramon Perez: ballot initiatives, I think Arizona is… has got potentially 30 or 40 that could end it. They might end up setting a record this year, possibly on the number of… and they do… they do in-person voting, right? So you gotta physically go to a polling location.
165
00:25:40.350 –> 00:25:53.930
Ramon Perez: have a line of people standing behind you, and now you’re reading about some bond issuance or some constitutional amendment that you’re seeing for the first time. I mean, it’s a lot of cognitive load for people, so that’s why they typically know who’s at the top of the ticket.
166
00:25:53.930 –> 00:26:04.989
Ramon Perez: But judges, city councilmen, you know, state reps, a lot of that gets left blank, or they just vote party lines. So, really, that’s part of the problem we’re trying to solve, and I think
167
00:26:04.990 –> 00:26:21.249
Ramon Perez: When the paper ballots go away, you actually find that you can have a much better, more comfortable voting experience where people feel like they had the opportunity to educate themselves in the comfort of their own homes before they make a pretty important decision.
168
00:26:21.630 –> 00:26:37.839
Jon Schwabish: Yeah, for sure. So, now, so we’re recording this in mid-April. A lot of the filing deadlines for the November elections are around this time, depending on state, you know, usually April 1, May 1, June 1.
169
00:26:38.040 –> 00:26:48.750
Jon Schwabish: Are you… I’m guessing you’re seeing, sort of, an uptick in interest. Are there other things that you’re thinking about trying to bring into the tool?
170
00:26:49.300 –> 00:26:51.729
Jon Schwabish: leading up to the November midterms?
171
00:26:52.810 –> 00:27:08.499
Ramon Perez: Right, the main thing that we want to do is, as… okay, so this year was the first year we’ve expanded to additional state legislatures, so we… all… Americans in all 50 states, if you’re registered to vote in any state, you can vote on federal bills right now.
172
00:27:08.500 –> 00:27:14.299
Ramon Perez: But this year was the first year we incorporated state legislation in additional states, including Virginia.
173
00:27:14.300 –> 00:27:21.680
Ramon Perez: Washington, Utah, Arizona, Michigan, Massachusetts. And so, those legislative sessions are about to come
174
00:27:21.680 –> 00:27:34.809
Ramon Perez: to an end, and so we will then produce the scorecards where we match the legislators, what they did in the session to… compared to what people wanted in the app. So, a lot of our main priority
175
00:27:34.870 –> 00:27:44.630
Ramon Perez: In prepping for election season is to provide people the best data we can provide them on how do their legislators vote, and what do their districts want,
176
00:27:44.630 –> 00:27:55.479
Ramon Perez: in the app compared to how their legislators voted, so they’ve got some concrete data to work with when they go to cast a ballot in November. I think that’s our primary mission. Inform people
177
00:27:55.520 –> 00:27:59.670
Ramon Perez: In a… a factually accurate, concrete way.
178
00:28:00.200 –> 00:28:03.540
Ramon Perez: Is the person who currently represents your district
179
00:28:03.810 –> 00:28:06.090
Ramon Perez: Doing their basic job, which is…
180
00:28:06.520 –> 00:28:19.790
Ramon Perez: go to Washington, or go to your state capitol, and represent that district. What did the district want, and what did that person do? So that is… that is our… that is our mission, and that’s what we’d be focused on as we ramp up for election season.
181
00:28:20.380 –> 00:28:20.990
Jon Schwabish: Great.
182
00:28:21.180 –> 00:28:32.910
Jon Schwabish: Just with the, with the, last few minutes we have left, we talked a lot about, like, additional things you could do, but looking past November into the next
183
00:28:33.040 –> 00:28:34.930
Jon Schwabish: Year, 5, 10 years.
184
00:28:35.060 –> 00:28:36.650
Jon Schwabish: what,
185
00:28:36.880 –> 00:28:49.729
Jon Schwabish: what are you thinking about, or what are you hoping the Digital Democracy Progress project, can accomplish, and what features can you add, and, like, what is that 5-10 year plan?
186
00:28:51.270 –> 00:28:53.720
Ramon Perez: you know, I think…
187
00:28:53.760 –> 00:29:13.270
Ramon Perez: it’s a double-edged sword with this technology, because the technology unlocks, like, infinite possibilities. So I have to… I have to try not to let my mind wander into every single thing that we can build. Right. Because, really, I think that’s so exciting. We are really, right now, at the… we have the ability to completely reimagine
188
00:29:13.270 –> 00:29:21.599
Ramon Perez: what American democracy looks like, and there’s a whole group of reformers and builders out there doing this kind of thing, but concretely for Digital Democracy Project.
189
00:29:21.610 –> 00:29:25.260
Ramon Perez: Our ambition is to get all 50 state legislatures
190
00:29:25.380 –> 00:29:30.300
Ramon Perez: into our system, so you… so any American can vote on any state-level legislation.
191
00:29:30.380 –> 00:29:44.579
Ramon Perez: In 2027. So that is our… that is our goal. We currently have 7 states, we want to be at 50 states. That’s very ambitious. Definitely members of my team who are saying that’s going to be too much too fast, but I actually think…
192
00:29:44.580 –> 00:29:56.759
Ramon Perez: This is what’s kind of exciting, is that, like, I am… I have been leaning in, really, with Cloud Code and code generation. I was using, Codex last year, but Claude Code is truly a game changer this year.
193
00:29:56.760 –> 00:30:06.129
Ramon Perez: And we’re… I think we’re going to be able to push so much faster than we could have before, and trying to get to all 50 states, to me, seems more…
194
00:30:06.130 –> 00:30:12.189
Ramon Perez: doable now, like this, like, in 2026 than it did, if you had asked me, even the last
195
00:30:12.480 –> 00:30:26.849
Ramon Perez: couple months of 2025. I mean, that’s the, like, the blinding speed of development is really remarkable, and it’s gonna allow us to build really cool features much, much faster. So, getting to all 50 states, and then after that,
196
00:30:27.250 –> 00:30:35.710
Ramon Perez: I have a goal, and we’ve always had it in our heads that we want to be at the local level. We want people to be able to weigh in on what’s happening in their city council.
197
00:30:35.710 –> 00:30:49.539
Ramon Perez: What’s happening in their county board? What’s happening in their school board? You know, some sort of rezoning goes on where your kid is now going to be put into a different school, and a lot of people are finding out about that after the fact, because if you’re not
198
00:30:49.850 –> 00:31:04.919
Ramon Perez: able to attend the school board meetings at, you know, 2 o’clock on a Tuesday, you don’t really hear about all the debate, but we have the tools now to bring the debate directly into people’s hands in an app on their phone, and so that’s where I want to be
199
00:31:04.920 –> 00:31:18.019
Ramon Perez: In the… in the out years, you know, after we get to all 50 state legislatures. But getting the data, you know, you’re a data guy, John, when… when you… we… we heavily lean on OpenStates, which is… which is a project
200
00:31:18.020 –> 00:31:28.769
Ramon Perez: That has been scraping the vote records of every legislator at the, from Congress and all 50 states, and they do incredible work of
201
00:31:28.930 –> 00:31:41.809
Ramon Perez: pulling recent copies of the bills, all the amendments, and then how did every legislator vote on every bill, which is not easy work. It’s a lot of data scraping of very complex, messy data on state government websites.
202
00:31:41.890 –> 00:31:54.240
Ramon Perez: to try to get down to the local level is just infinitely harder. You go from 50 states to 10,000 municipalities, 3,000 school boards. It’s a… it’s a challenge, but I happen to think that
203
00:31:54.610 –> 00:32:11.620
Ramon Perez: AI-generated browser, automation and Claude Code is gonna… is gonna open up the possibility for us to do local-level, data collection of how did every city councilman in Walla Walla, Washington vote.
204
00:32:11.620 –> 00:32:11.940
Jon Schwabish: Yeah.
205
00:32:11.940 –> 00:32:17.959
Ramon Perez: on a particular housing bill, or something like that. I think… I think we can get there because of what AI gives us.
206
00:32:18.380 –> 00:32:27.329
Jon Schwabish: Wow, yeah, that’s great. Okay, so, folks have listened to this, they’re really interested, where can they find you, where can they find the tool, where can they get the app?
207
00:32:27.860 –> 00:32:42.049
Ramon Perez: You can visit us at digitaldemocracyproject.org. Our partners are Votes, V-O-A-T-Z. You can go to Votes.com. You can download the Votes app, and our partnership with Votes is that anybody who downloads the app gets to see bills
208
00:32:42.300 –> 00:33:02.069
Ramon Perez: populating in the app. So, if you download the Votes app, and you go through the voter verification with a photo ID, and you’re on the voter file, we can find you on the voter file. You’ll see legislation start to populate. If you are in one of the seven states where we have state legislation, you’ll see federal and state. If you’re not in one of those seven states, you’ll see federal bills
209
00:33:02.070 –> 00:33:10.869
Ramon Perez: But you’ll see… you’ll have something you can participate on right away. So you can download the Votes app, you can go to digitaldemocracyproject.org, and you can see the maps
210
00:33:10.870 –> 00:33:23.210
Ramon Perez: Where we show congressional districts and how people are voting. And, you can also talk to our chatbot on there and get really detailed questions answered about policy. So yeah, give us a look.
211
00:33:23.520 –> 00:33:29.819
Jon Schwabish: Awesome. Thanks, Ramon. Appreciate it. This is really, really interesting stuff, really interesting work, and yeah, thanks for coming on the show.
212
00:33:30.180 –> 00:33:31.879
Ramon Perez: Thank you, John, for having me, I appreciate it.