New Study Unveils the Total Cost to Attack Bitcoin and Ethereum

February 16, 2024 Bitcoin.com

Cryptocurrency charities can exploit the ‘gambler’s fallacy’ to reap larger donations — study

November 9, 2023 Coin Telegraph

Charity

According to the researchers, people’s penchant to misinterpret patterns could be a boon for charitable organizations that accept cryptocurrency.

A team of academic researchers from the U.S. recently published a study exploring how the “gambler’s fallacy” affected cryptocurrency donations. Their findings indicate that organizations accepting crypto donations could benefit from timing the market.

Essentially, the team’s work explores the idea that people generally misinterpret certain pattern signals when it comes to finance. Charities that understand the penchant for crypto holders to hold or move assets based on perceived market conditions may be able to optimize their strategies to reap larger donations.

Per the paper:

“Our findings support actionable recommendations for how charities can design more intentional fundraising campaigns to take advantage of the cost and time efficiencies of cryptocurrencies. By considering recent changes in cryptocurrency prices and highlighting the urgency to donate, charities can design more effective strategies to engage cryptocurrency donors.”

The team tested their premise through an empirical study of cryptocurrency donations to 117 campaigns at an online crowdfunding platform. They also conducted a controlled online experiment studying features of cryptocurrency donation context.

After careful analysis, the team determined that market movement was directly correlated to donation "activation" (first time donations) and donation sizes.

According to the paper, the online experiment expanded on the empirical analysis and demonstrated that “donors’ decisions are affected by recent changes in asset price, consistent with the gambler’s fallacy heuristic.”

The gambler’s fallacy, also commonly called the Monte Carlo fallacy, refers to the tendency for people to misinterpret statistically meaningless historical events, such as the flip of a coin, as a predictor for future odds.

As an example of the gambler’s fallacy, if a person flips a coin 10,000 times in a row, and it lands on heads each time, an observer might think that the next coinflip has a higher chance of landing on tails because, as the above video explains, “it’s due.”

In reality, the odds of a coin landing on heads or tails is always exactly one-in-two with no regard for historical outcomes.

During the study, the researchers determined that participants are more likely to be activated to donate after experiencing declines in asset value. This purportedly occurs because donors feel more confident that prices will go up after their donation due to the gambler’s fallacy. “Moreover,” the paper continues, “we observe that participants’ reliance on the gambler’s fallacy is amplified when they face urgent donation appeals.”

Ultimately, the paper concludes that these insights could be used as empirical evidence in the decision-making process for organizations and individuals managing charities that accept cryptocurrency donations.

Anthropic built a democratic AI chatbot by letting users vote for its values

October 18, 2023 Coin Telegraph

The value responses from 1,000 test subjects were used to tune a more democratic large language model.

In what may be a first of its kind study, artificial intelligence (AI) firm Anthropic has developed a large language model (LLM) that’s been fine-tuned for value judgments by its user community.

What does it mean for AI development to be more democratic? To find out, we partnered with @collect_intel to use @usepolis to curate an AI constitution based on the opinions of ~1000 Americans. Then we trained a model against it using Constitutional AI. pic.twitter.com/ZKaXw5K9sU
— Anthropic (@AnthropicAI) October 17, 2023

Many public-facing LLMs have been developed with guardrails — encoded instructions dictating specific behavior — in place in an attempt to limit unwanted outputs. Anthropic’s Claude and OpenAI’s ChatGPT, for example, typically give users a canned safety response to output requests related to violent or controversial topics.

However, as innumerable pundits have pointed out, guardrails and other interventional techniques can serve to rob users of their agency. What’s considered acceptable isn’t always useful, and what’s considered useful isn’t always acceptable. And definitions for morality or value-based judgments can vary between cultures, populaces, and periods of time.

One possible remedy to this is to allow users to dictate value alignment for AI models. Anthropic’s “Collective Constitutional AI” experiment is a stab at this “messy challenge.”

Anthropic, in collaboration with Polis and Collective Intelligence Project, tapped 1,000 users across diverse demographics and asked them to answer a series of questions via polling.

The challenge centers around allowing users the agency to determine what’s appropriate without exposing them to inappropriate outputs. This involved soliciting user values and then implementing those ideas into a model that’s already been trained.

Anthropic uses a method called “Constitutional AI” to direct its efforts at tuning LLMs for safety and usefulness. Essentially, this involves giving the model a list of rules it must abide by and then training it to implement those rules throughout its process, much like a constitution serves as the core document for governance in many nations.

In the Collective Constitutional AI experiment, Anthropic attempted to integrate group-based feedback into the model’s constitution. The results, according to a blog post from Anthropic, appear to have been a scientific success in that it illuminated further challenges towards achieving the goal of allowing the users of an LLM product to determine their collective values.

One of the difficulties the team had to overcome was coming up with a novel method for the benchmarking process. As this experiment appears to be the first of its kind, and it relies on Anthropic’s Constitutional AI methodology, there isn’t an established test for comparing base models to those tuned with crowd-sourced values.

Ultimately, it appears as though the model that implemented data resulting from user polling feedback outperformed the base model “slightly” in the area of biased outputs.

Per the blog post:

“More than the resulting model, we’re excited about the process. We believe that this may be one of the first instances in which members of the public have, as a group, intentionally directed the behavior of a large language model. We hope that communities around the world will build on techniques like this to train culturally- and context-specific models that serve their needs.”

US voters across 4 swing states oppose anti-crypto pres candidates: Coinbase

September 20, 2023 Coin Telegraph

campaign

A poll of voters in New Hampshire, Nevada, Ohio and Pennsylvania found 55% would be less likely to vote for anti-web3 presidential candidate, said Coinbase.

Crypto exchange Coinbase says it will focus its "Stand with Crypto" campaign on nine states in the U.S., including four “swing states” with voters polled as less likely to choose anti-crypto presidential candidates.

The four “swing states” include New Hampshire, Nevada, Ohio, and Pennsylvania, which are just some of the states the crypto exchange is set to focus its “Stand with Crypto” campaign, it said in a Sept. 19 blog.

“Polling in the fall of 2022 showed that in the key states of NH, NV, OH, and PA, over half (55%) of voters stated that they would be less likely to vote for candidates who oppose crypto and Web3,” said Coinbase, citing survey data collected in a Morning Consult poll a year prior.

52M Americans own crypto. They’re younger & more diverse than the US population in total. But their voices aren’t being heard by Congress.

That’s about to change.

Here’s what we’re doing—and what you can do to help mobilize 1M people to #StandWithCrypto: https://t.co/stmH0WVUhO
— Coinbase ️ (@coinbase) September 19, 2023

The October-published poll included 800 likely voters across the four states. Between 13-19% of respondents in the four states claimed to own cryptocurrency at the time of the polling.

*Percentage of voters in the four swing states that own cryptocurrencies or nonfungible tokens. Source: Morning Consult*

Coinbase also cited that in Nevada, Ohio and Pennsylvania in particular, more than 40% of crypto owners use blockchain to remit money overseas to help family members pay for food, housing and health care for far less than what banks typically charge with international wire transfers.

Coinbase attempts to mobilize crypto users

The highlighted research comes amid an ongoing “Stand With Crypto” campaign, a 14-month-long campaign launched by Coinbase in August, pushing for crypto legislation in the country.

Coinbase said the campaign will focus on the four “swing states” but will also include an intense focus on Arizona, California, Georgia, Illinois, and Wisconsin, which have an “over-index” when it comes to crypto owners, it said.

“While a growing number of elected officials in Congress are advocating for legislation to regulate crypto and drive progress in the financial system, others are simply choosing to protect the status quo.

Coinbase is also set to organize a “Stand with Crypto Day” will take place in Washington D.C. on Sept. 27 to advocate for better cryptocurrency innovation and policy.

The campaign is also encouraging crypto owners and supporters to take a moment to call their members of Congress and ask them to pass clear, sensible legislation.

“Today we kick off this effort by rallying the collective energy of the community, and taking the fight off X (formerly known as Twitter) and on to the phones."

“The campaign will encourage crypto owners and supporters to take one minute of their day to call their member of Congress and ask them to pass clear, sensible legislation,” it added.

Magazine: Binance, Coinbase head to court, and the SEC labels 67 crypto-securities: Hodler’s Digest, June 4-10

Elon Musk tweets and Twitter bot spam influences altcoin prices: Study

August 3, 2023 Coin Telegraph

Altcoins

The study also questioned whether FTX or Alameda had any role in coordinating Twitter bot activity for its own gain.

Crypto-spouting Twitter bots could be playing a much larger role in artificially inflating the price of altcoins than previously understood, a new study has suggested.

Using a sample of various FTX-listed cryptocurrencies in a study published Aug. 2, the Network Contagion Research Institute (NCRI) said it analyzed over 3 million tweets posted between Jan. 1, 2019, to Jan. 27, 2023, surrounding 18 altcoins.

The study found that Twitter bot activity played a crucial role in amplifying the value of these cryptocurrencies, including The Sandbox (SAND), Gala (GALA), Gods Unchained (GODS) and LooksRare (LOOKS), with half of the coins showing signs of price influence as a result of tweet bot activity.

It also found that these inauthentic tweets would increase after FTX posted about the token on social media, which it said raises questions about whether FTX or Alameda Research could have played a role in coordinating the bot activity.

*Each line shows the average number of botted tweets, which increased following FTX’s listing and mention. Source: NCRI*

"In fact, for half of the FTX listed coins in the sample, inauthentic tweet volume showed signs of forecasting subsequent price. This suggests that inauthentic networks successfully and deliberately deployed to influence changes in FTX coin prices," it said, adding:

“It begs the question, did FTX or Alameda engage in coordinated inauthentic activity on social media to artificially inflate market values?”

Musk’s tweets impact PSYOP and PEPE

The study also looked into the impact of bot activity and Elon Musk's crypto-adjacent tweets on two recent memecoins, suggesting the prices of Pepe (PEPE) and PSYOP have been influenced by both these factors.

NCRI detected a surge of newly created bot accounts before the launch of PEPE — which all went on to tweet about one of the two coins.

Pepe Coin and PSYOP leveraged memes and were also boosted by two of Musk’s tweets that seemingly gave a nod to each of the tokens, the study said.

Musk’s May 13 tweet of a Pepe meme caused the token's price to jump over 50% within 24 hours.

Explaining Constitutional Amendments pic.twitter.com/oYkMPBe9Zi
— Elon Musk (@elonmusk) May 13, 2023

Alongside Musk’s tweets, account creation surges took place a day prior to Pepe’s April 17 launch which suggested an orchestrated effort to use bots to amplify the token’s popularity.

*Chart showing a significant surge in bot accounts a day before Pepe’s launch. Source: NCRI*

The study’s researchers said this phenomenon could also affect stocks and other securities. They pointed to the social media frenzy in 2022 surrounding so-called “meme stocks” such as Gamestop and AMC.

Magazine: How smart people invest in dumb memecoins — 3-point plan for success

Token hoarders defeat the purpose of most DAOs: Study

July 14, 2023 Coin Telegraph

Coin Telegraph

The study also showed that decentralized organizations work best when they’re built around a tight-knit group of focused participants.

A pair of researchers from the University of Texas at Austin and Princeton University conducted a study to determine how tokenization affects decentralization in decentralized autonomous organizations (DAOs). Their findings indicate that many of the challenges to autonomy are related to the reasons individual users have for participating.

According to their research, the larger a DAO grows, the more incentive participants have to consider DAO tokens investments:

“The presence of investors diverts the subsidy away from users and thus harms their participation. More importantly, investors may even take a majority stake to seize control of the platform.”

In a typical token-based DAO scheme, rather than having a CEO or leader implement decisions, individual participants are issued distributed authority through tokens. This prevents the people who maintain the DAO from exploiting the participants because their tokens work like votes.

“The key distinction between tokens and securities,” per the team’s research paper, “is that tokens are a claim to the platform’s services while securities are a claim to its revenue.”

As long as the participants in a DAO are aligned in purpose and willing to spend their tokens to vote for actions that move that purpose forward or on services and utilities that provide value to the community, the DAO tends to thrive.

The researchers modeled DAOs over time to determine how user growth and tokenization affect outcomes. Their primary finding, according to the team’s research paper, is that tokenization serves the purpose of shifting ownership from initial equity holders to a platform’s users, but the tradeoff is that there’s no single entity that can subsidize network participation.

This evidently leaves the gates open for investors to treat purpose-driven DAOs like traditional stocks.

“The ability to get high returns has hurt cryptocurrencies as mediums for payment, because people don’t want to spend it,” said lead researcher Michael Sockin in a press release, “They can easily take us back to being like Amazons and Apples, which is the whole issue we were trying to move away from."

Social media discussions play a crucial role in influencing crypto returns: Study

July 10, 2023 Coin Telegraph

Coin Telegraph

The researchers also determined that “news sentiment” is a much less effective predictor of cryptocurrency returns.

Researchers at Pennsylvania State University recently analyzed whether attitudes and emotionality surrounding cryptocurrency could help predict returns. What they found may stand in stark contrast to related financial markets.

According to the team’s research paper, social media plays an outsized role in adoption and activity rates, while cryptocurrency journalism isn’t a great predictor of market movement:

“Our findings indicate that social media sentiment significantly predicts crypto returns, while sentiment from news media does not.”

The researchers used natural language processing to analyze millions of financial news articles and social media comments and generated sentiment scores along 53 topics and attention metrics for over 300 cryptocurrencies.

They then compared the ground truth returns over a given period of time to the coinciding news and social media sentiment.

*Chart of social media attention and crypto market cap. Source: "An Anatomy of Cryptocurrency Sentiment"*

Perhaps most interesting among their findings is their conclusion that while social media sentiment is a good predictor of crypto returns, the risk premium channel is not.

The risk premium channel is a sort of lens by which consumers make investment decisions. It is directly related to market and asset volatility.

Cryptocurrency is often discussed as a highly volatile asset. In typical markets, such volatility usually leads to a higher risk premium and lower adoption and activity.

Taking the housing market as an example, research shows that as market volatility increases, consumer sentiment decreases and would-be purchasers tend to become risk-averse.

The Penn State team’s research indicates that this isn’t the case with cryptocurrency. In its conclusion, the team writes that market exuberance is positively related to momentum but that it “does not positively predict volatility.”

“This suggests,” the paper continues, “that sentiment influences returns through price perception and demand shocks rather than the risk premium channel.”

The researchers ultimately conclude that this may be due to the large number of consumer investors with large cryptocurrency portfolios active on crypto social. They also suggest that further research on the relationship between social media sentiment and crypto returns is merited.

Study Reveals Top Countries Fueling Meme Coin Interest in 2023

May 8, 2023 Bitcoin.com

Over half of Americans fear ‘major impact’ by AI on workers: Survey

April 21, 2023 Coin Telegraph

Bias

More respondents said AI will “hurt” American workers more than it will “help” them over the next 20 years.

Nearly two-thirds (62%) of Americans think implementing artificial intelligence (AI) in the workplace will have a “major impact” on American workers within the next 20 years, leaving many employees “wary” and “worried” about what their future holds.

An April 20 Pew Research report found 56% of the 11,004 adults surveyed in the United States said AI will have a major impact on the U.S. economy too. Another 22% believed AI will impact the economy to a minor degree.

Only 13% of participants believed “AI will help more than hurt” American workers, whereas 32% thought the opposite. The rest of the participants predicted “AI will equally help and hurt” American employees (32%) or were unsure (22%).

While 62% of Americans believe AI will have a major impact on workers generally, just 28% believe its use in the workplace will have a major effect on them personally. https://t.co/ExpDNlOsWl pic.twitter.com/biMSX1AUln
— Pew Research Center (@pewresearch) April 20, 2023

The study didn’t directly ask participants whether they thought they would lose employment to AI but many respondents cited worry that an AI-enabled workplace would lead to increased surveillance, data mismanagement and misinterpretations.

Pew Research said there is a “consensus” that many American workers feel like they would be watched “Big Brother” style, with 81% citing the concern.

71% of respondents said they oppose the idea of AI being used to help make a final decision in the hiring process.

Nearly two-thirds said they would be most bothered by AI tracking their minute-to-minute movements, and around half cited potential frustrations around an AI keeping track of how many hours they’re at their desk and recording exactly what they’re working on.

*For every participant that was in favor of AI being used in the hiring process, 10 opposed it. Source: Pew Research*

Just under 40% cited concern that AI would be used to evaluate their performance.

Despite the mixed views on what AI would offer to the workforce, two-thirds of respondents said they wouldn’t want to apply for a job where AI was used to make hiring decisions.

One surveyed man in his 60s explained that AI shouldn’t be used for that purpose because it can’t judge character:

“AI can’t factor in the unquantifiable intangibles that make someone a good co-worker ... or a bad co-worker. Personality traits like patience, compassion and kindness would be overlooked or undervalued.”

“It’s a ‘garbage in, garbage out’ problem,” another surveyed woman explained.

Not everyone agreed though as a man in his 50s explained AI has the potential to fill the shoes of a hiring manager:

“I think the AI would be able to evaluate all my skills and experience in their entirety where a human may focus just on what the job requires. The AI would see beyond the present and see my potential over time.”

Just under half of the participants said AI would treat all applicants in the same way “better” than what hiring managers do, while 15% said AI would be “worse.” Under 15% said the treatment would be “about the same.”

Those surveyed who claimed AI would lead to “better” treatment explained the technology would help circumvent biases and discrimination based on age, gender and race.

Others believed AI may reinforce the same prejudices that companies are trying to eradicate.

The motivation to carry out the study was partly prompted by what Pew Research describes as the “rapid rise of ChatGPT” — an AI chatbot released by OpenAI on Nov. 30.

Magazine: NFT Creator, Emily Xie: Creating ‘organic’ generative art from robotic algorithms