Dall-E Mini, the AI-powered text-to-image generator has taken over the internet. With its ability to render nearly anything your meme-loving heart desires, anyone can make their dreams come true.
DALL-E 2, a portmanteau of Salvador Dali, the surrealist and Wall-E, the Pixar robot, was created by OpenAI and is not widely available; it creates far cleaner imagery and was recently used to launch Cosmpolitan’s first AI-generated cover. The art world has been one of the first industries to truly embrace AI.
The open-sourced miniature version is what’s responsible for the memes. Programmer Boris Dayma wants to make AI more accessible; he built the Dall-E Mini program as part of a competition held by Google and an AI community called Hugging Face.
And with great technology, comes great memes. Typing a short phrase into Dall-E Mini will manifest 9 different amalgamations, theoretically shaping into reality the strange images you’ve conjured. Its popularity leads to too much traffic, often resulting in an error that can be fixed by refreshing the page or trying again later.
If you want to be a part of the creation of AI-powered engines, it all starts with code. CodeAcademy explains that Dall-E Mini is a seq2seq model, “typically used in natural language processing (NLP) for things like translation and conversational modeling.” CodeAcademy’s Text Generation course will teach you how to utilize seq2seq, but they also offer opportunities to learn 14+ coding languages at your own pace.
You can choose the Machine Learning Specialist career path if you want to become a Data Scientist who develops these types of programs, but you can also choose courses by language, subject (what is cybersecurity?) or even skill - build a website with HTML, CSS, and more.
CodeAcademy offers many classes for free as well as a free trial; it’s an invaluable resource for giving people of all experience levels the fundamentals they need to build the world they want to see.
As for Dall-E Mini, while some have opted to create beauty, most have opted for memes. Here are some of the internet’s favorites:
no fuck every other dall-e image ive made this one is the best yet pic.twitter.com/iuFNm4UTUM
— bri (@takoyamas) June 10, 2022
There’s no looking back now, not once you’ve seen Pugachu; artificial intelligence is here to stay.
The Census has one goal: "Counting everyone once, only once and in the right place." Every 10 years, the US Census Bureau conducts a survey of every single person living in the United States.
The United States has been conducting census surveys since the 1790's because when the founding fathers were building their fledgling democracy they decided that population would be the basis of political power.
The 2020 Census asks a few simple questions about you and everyone who was living with you as of April 1, 2020. The survey asks about the number of people living in your household, and each persons age, sex, and race.
The September 30th deadline is rapidly approaching, but some American's still haven't filled out their census. Not filling it out, has some very real consequences. Here are six reasons why you should definitely fill out your census immediately.
1. It's illegal not to.
Your response is required by law.Getty Images
Filling out the census is mandatory, and everyone living in the U.S. and its five territories as of April 1, 2020, is supposed to be counted. This includes children, babies, people without homes, college students, and immigrants regardless of their legal status.
According to United States Code, Title 13 (Census), Chapter 7 (Offenses and Penalties), SubChapter II, if you're over 18 and refuse to answer all or part of the census, you can be fined up to $100. If you give false answers, you're subject to a fine of up to $500. If you offer suggestions or information with the "intent to cause inaccurate enumeration of population," you are subject to a fine of up to $1,000, up to a year in prison, or both.
2. If you don't, someone will show up at your house.
Census Bureau EnumeratorU.S. Census Bureau
If you haven't already filled out the census, you can expect a visit from a US Census taker as part of NRFU. The Nonresponse Followup operation (NRFU) is when census takers visit nonresponding households to survey their inhabitants in-person and enter their answers on their secure Census Bureau smartphone.
If no one is home when the census taker visits, they will leave a notice of their visit with information about how to respond online or by phone. As necessary, they will make additional visits to collect responses from the household.
NRFU is a huge operation, this year they are expected to send enumerators to approximately 56.4 million households that have not responded to the 2020 Census.
3. It affects your representation
Apportionment map of 2010 CensusU.S. Census Bureau
The census determines how many representatives each state will have in Congress for the next 10 years. The US House of representatives has 435 seats, each of which is allocated based on population. For instance, Texas gained four seats after the last Census, while New York and Ohio lost two seats each. State and local officials also use census results to help redraw congressional, state, and local district boundaries.
4. It affects your community's funding.
Billions in government funding are allocated through census data every year. Andy Warhol
The Census also affects funding. Each year, approximately $675 billion in federal funds is spent on schools, hospitals, roads, public works and other vital programs, and this money is divided up largely based on population. Responding to the census helps your community receive its fair share of that funding.
Census data isn't only used by the public sector. It is also used by many businesses in the private sector to decide where to build factories, offices, and stores, and developers use the census to decide where to build new homes.
Local governments use the census so they can plan for every emergency from riots to fires to hurricanes, and advocacy groups use the census to support community initiatives involving legislation, quality-of-life and consumer rights.
5. It's so easy and safe!
This is the first year you can fill out your census online!U.S. Census Bureau
This year, there are three different easy ways to fill out your census.
- By mail. You should have received a paper survey in the mail in April, which can still be mailed back to the U.S. Census Bureau.
- Online. This is the first time the census can be completed online! Visit www.2020census.gov to fill out the survey. The online survey has 13 available language options.
- By phone. Respondents can call the Census Bureau at 844-330-2020, and can be provided with answers over the phone in 13 languages.
You may be thinking that you don't care how easy it is, and no matter what, you don't want to share your information with the government. But the census really is a safe and secure way to participate in democracy. It is strictly against the law for the Census Bureau to disclose or publish any census information that identifies an individual. Title 13 makes it very clear that the data collected can only be used for statistical purposes—and it is not allowed to be used for anything else, including law enforcement. No law enforcement agency can access or use your personal information at any time, including the Department of Homeland Security, ICE, the FBI, and the CIA.
6. COVID-19 might lead to an undercount, which hurts vulnerable populations.
Undercounts hurt everyone, but they hurt vulnerable populations the most.Getty Images
The coronavirus has changed a lot about how the Census operates. In-person efforts to reach census respondents have been significantly reduced and delayed, the Census office had to undergo a month long hiring freeze right as they began sending out survey, and the deadline for filling it out has shifted twice.
The NRFU operation was originally scheduled for May 13 through July 31, 2020, but these dates have been adjusted due to the COVID-19 pandemic and are now set to take place between August 11 and September 30, 2020. This reduction from two and a half months to one and a half months has caused some concern.
The Census Bureau is also having trouble hiring and keeping employees during the pandemic. They say they require 435,000 census-takers to complete their accelerated count, but as of Aug. 18 they had onboarded just 309,000. A report from the Government Accountability Office explains this discrepancy may be due to high attrition rates. The Census Bureau expected that about 10% of employees that started training would not show up for field work, but surprisingly the rate was closer to 35%.
The Bureau's original COVID-19 plan was to extend the census count's follow-up period from July 31 to October 31. Given this extension,the Bureau also requested that Congress delay the mandated reporting of census results to the president from December 31, 2020 to April 30, 2021. However, the extension was never approved and on August 3, the Bureau announced its intention to end all follow-up activities by September 30 and report the results to the president by December 31.
The Census Bureau says they are on track to meet this goal, but many have criticized the decision and blamed it on pressure from the White House. Four former Census Bureau directors who served under nine presidents have made a statement saying that without the full deadline extension, "The Census Bureau will not be able to carry out the NRFU fully and will be forced to take steps such as fewer in-person visits and rely instead on the use of administrative records and statistical techniques on a much larger scale than in previous census." This could result in a serious undercount of hard-to-count populations.
Hard-to-count populations have always taken extra efforts to reach. These populations include the homeless, residents of dormitories or group homes, racial minorities, immigrants, rural residents, and residents on Native American reservations. Some groups are hard to count simply because they are hard to locate, like homeless people or people in remote locations, but others don't respond on purpose because they don't trust the government.
Unfortunately, the populations that are most likely to go uncounted are also those who need funding and representation the most. These populations are often already underserved by the government and an undercount could affect the next 10 years of resources available to them. Homeless people, minorities, immigrants, and rural communities are all at risk of being underrepresented in legislative bodies, and they also risk receiving less funding for desperately needed social programs, hospitals, and schools.
September 30 is the last day to be counted in the United States Census. As of September 1st, approximately 64.9% of the country has already self-responded, and another 17.9% have been counted through follow-up efforts by the bureau. Unfortunately, that still leaves approximately 17.6% of the population to locate and count in the next 30 days. Make sure your community gets the resources and representation it deserves by filling out the Census today
For more well-researched, unbiased information on today's biggest issues, follow Alexandra's Instagram account The Factivists.
Statistical anomalies in previous elections point to frightening vulnerabilities in how votes are tallied
In 2012, Mitt Romney was declared the winner of the GOP's Iowa caucus by a margin of just eight votes.
That result was announced at 1:30 AM on Caucus night, but two weeks later there was a different result. The Republican Party of Iowa had performed a recount of the votes and Rick Santorum—a candidate broadly considered too far outside the mainstream to stand a chance in the general election—was revealed as the actual winner. By that point of course it was far too late. Romney had already gotten the media attention that comes with winning, and had capitalized on that sense of momentum to achieve a solid victory in the New Hampshire primary.
If that sounds familiar, that's because there is currently a "recanvassing" under way in Iowa to reassess the results of a contest that will once again be decided by a very slim margin. Though Bernie Sanders managed a clear victory in the popular vote, Iowa's elaborate system for awarding "State Delegate Equivalents" at each caucus location (to then be converted to the actual delegate count that determines the Democratic party's nominee), has resulted in a near-tie between Senator Sanders and former South Bend Mayor Pete Buttigieg.
The official winner may not be announced for days or weeks—or it may never be known at all. In the mean time, Pete Buttigieg declared himself the winner on Caucus night, and has been treated as such by much of the media. With Sanders looking more and more like the frontrunner, Buttigieg is seen as much more palatable to moderate general election voters, and many powerful donors and party insiders would much rather he get the nomination. With that in mind, the chaos in Iowa—particularly the faulty app and the release of partial results that seemed to favor Buttigieg—has already sparked speculation of party corruption and election rigging for Sanders voters holding on to memories of the DNC's favoritism toward Hillary Clinton in 2016, but it may be more instructive to consider the model of the 2012 Republican primaries.
2012 was not a simple year for Republicans in the way 2016 was for Democrats. Much like the 2020 field of Democratic candidates, there was not a presumptive nominee, but rather a wide field of contenders with centrists—Mitt Romney and Jon Huntsman—struggling against the enthusiastic support for more extreme candidates like Rick Santorum, Newt Gingrich, and Ron Paul. Polling seemed to swing wildly, from one candidate to the next until—according to some accounts—the RNC tipped the scales toward the man considered the most electable (read: bland and inoffensive) candidate.
The case of bias for Romney is as elaborate and detailed as every budding conspiracy theory about Acronym, Shadow Inc., and the new "Voter Protection Director" for the Nevada State Democratic Party—especially considering the contentious events of the Nevada Democratic Party state convention in 2016. Likewise, the overwhelming variety of those details is evidence—depending on your perspective—of either how insidious the whole plot was, or of the depths of its adherents' delusion.
For a start, there were a variety of issues in that year involving states attempting to increase their influence on the process by ignoring the traditional schedule for primaries and caucuses, skipping ahead of other states. The RNC was conflicted about how to handle that tactic, but many voters felt that the resulting schedule favored Romney's candidacy by allowing states where he polled well to vote earlier, further emphasizing the sense of building momentum. That much was acknowledged as a problem even at the time, but there are other issues that remain murkier. Was the miscount in Iowa intentional? Did the RNC combine fundraising with Romney's campaign too early? Did they pad Romney's delegate count to prevent a brokered convention? Most worryingly of all, were votes flipped to Romney in state primaries that lacked a paper trail?
Whatever the confusion in Iowa, the caucus system is at least public and relatively easy to monitor. But primaries that are carried out with all electronic voting machines are a black box, and the state parties run the show with little oversight. Unlike a federal election, primaries are fundamentally under the purview of political parties. They can choose their nominees however they like. They have chosen a roughly democratic system for a variety of reasons (to build enthusiasm, test candidates' campaign skills, and avoid voter alienation) but if they wanted to undermine the integrity of that process in order to ordain the candidate they see as standing the best chance in the general election, there would be little to stop them. According to two statisticians, Francois Choquette and James Johnson, that is exactly what happened in 2012.
Choquette and Johnson, 2012
When Choquette and Johnson analyzed vote totals out of hundreds of precincts, they discovered a strong tendency for voters in larger precincts to favor Mitt Romney more than did voters in smaller precincts. These results have been criticized as attributable to demographic differences between precincts, but when researchers looked at results out of precincts that kept paper records of voting, that tendency disappeared. Likewise in Utah—where Romney was always expected to win by a wide margin—the results showed no shift in preference toward Romney based on the size of the precinct. The trend was so distinct in competitive precincts with no paper trail that Choquette and Johnson were unable to account for it as a result of chance or any factor other than deliberate fraud. According to their work, Romney received approximately a 7% bump in the most populous precincts in multiple states as a result of votes flipped from other candidates—allowing him to secure the nomination handily.
Choquette and Johnson's evidence was brought to court in a lawsuit filed by a third statistician at Witchita State University. Beth Clarkson works in the university's National Center for Advanced Materials Performance, and was inspired by Choquette and Johnson's research to investigate similar anomalies in Kansas's 2014 general election—larger precincts trending toward Republican candidates. She sued to gain access to more detailed records in order to build a statistical model that could shed light on the question of fraudulent vote flipping. Unfortunately Clarkson's efforts were stymied by Kansas Secretary of State Kris Kobach, who argued that releasing the time-stamped records could somehow violate voter anonymity.
Kris Kobach with Donald Trump
You may recognize Kris Kobach as perhaps the greatest enemy of unbiased elections in recent American history. Kobach is the man whose voter fraud commission pushed the false narrative of millions of fake voters in support of discriminatory voter ID laws, and whose apparent political and white-supremacist motivation for adding a citizenship question to the national census resulted in the Supreme Court rejecting the change—which would have led to dramatic under-counts of immigrant populations, and a shift in congressional districts that would disproportionately benefit the Republican party. While he was able to halt Clarkson investigation, many have credited her lawsuit as being instrumental in the push for recently enacted legislation which now requires post-election audits in Kansas. Reached for Comment, Clarkson admitted that this was a step in the right direction, though she was "not impressed" with the audit techniques being utilized.
Overall, Clarkson seems pessimistic about the fidelity and security of our elections, saying that there is still a lot of potential for fraudulent vote counts, "anytime there's no way to check a paper record." This does include several races in the Democratic primaries, though the New Hampshire primary taking place today is utilizing paper ballots, which leave less room for tampering. Nationally, there has been a push to move toward that model, but for those of us who will be casting votes on electronic machines, Clarkson advocated that, where possible, voters check their electronic vote against the paper record to ensure their vote was recorded correctly. And for all voters, Clarkson had a reminder to check your voter registration online in advance of every election. With recent voter purges Clarkson says there have been many cases "of people being surprised when they arrive at the polls to vote and find out, 'Oh, they don't have me down as a registered voter.'"
Beyond that, electronic voting remains so opaque and vulnerable, all we can do is continue pushing for paper-based voting systems and remind our nation's political institutions—as the primary process continues through New Hampshire and beyond—that we are watching them; that we will not take any irregularities lightly.