Crypto VC market flashes green amid macroeconomic recession alarms

June 20, 2023 Coin Telegraph

May 2023 marked the second consecutive month of crypto VC activity growth, surpassing $1 billion in funding for the first time since September 2022.

Crypto venture capital investments were on the rise for the second month in a row in May despite the generally declining economic backdrop. Funding amounts surged 34% from April, and the number of individual deals jumped 62%, according to data from Cointelegraph Research’s Venture Capital Database.

Though inflation in the United States cooled from 4.9% in April to 4% in May — down from 9.1% in the summer of 2022 — the U.S. Federal Reserve still raised interest rates 10 consecutive times. Decreasing inflation tends to build trust among investors that inflation is controllable and that Federal Reserve measures will become softer, but the market is still in the waiting phase.

Purchase access to the Cointelegraph Research VC Database.

On June 14, the Fed announced it would pause interest rate hikes, which may become a bullish signal for financial markets, with crypto no exception. Cointelegraph Research’s Venture Capital Database reveals that the crypto VC market saw $1.1 billion in investments in May, the first month to surpass the $1 billion mark since September 2022, with June set to serve as a crucial benchmark for continued growth in VC investment trends.

Blockchain infrastructure is still on top

Breaking down May’s deals, the infrastructure sector still leads the market in capital inflows with $783.9 million in 23 rounds, over 68% of the total invested money. In terms of the number of deals, Web3 is on top with 24 deals conducted, but seeing only $170.1 million in funding. Decentralized finance lost ground in May, with 20 deals and $93.6 million raised. Centralized finance was not attractive for VCs, having only two deals at $24.8 million in total.

The top raisers in May included infrastructure solutions developers Worldcoin and Auradine alongside Web3 project Magic. Worldcoin’s $115 million Series C round saw the participation of Spark Capital, Zoom Ventures, Sound Ventures, Salesforce Ventures, Menlo Ventures and Google and was aimed at promoting World App, its custodial solution, and World ID, its decentralized identity solution.

Blockchain privacy and security provider Auradine raised $81 million in a Series A round with the backing of Marathon Digital Holdings, Celesta, Mayfield, Cota Capital and DCVC to promote the “next-generation web infrastructure” with artificial intelligence and zero-knowledge-proof solutions.

Web3 development and tooling project Magic got a $52 million Series B deal with PayPal Ventures and Volt Capital as backers, among others. The funds are intended to expand the company’s integration in European and Asia-Pacific markets.

As of June, the Fed’s streak of 10 consecutive interest rate hikes has ended. That may turn investment strategies back to risk-on, as short-term adjustments to interest rates significantly impact how capital is invested in both the traditional and decentralized finance spaces. However, VC activity is a lagging indicator and may experience tailwinds in the background of the news. To keep on top of VC activity, follow the Cointelegraph Research VC Database, which is updated weekly and tracks over 6,000 deals from 2012 through the present day.

The opinions expressed in this article are for general informational purposes only and are not intended to provide specific advice or recommendations for any individual or on any specific security or investment product.

Meta’s new ‘Voicebox’ AI is a text-to-speech tool that learns like ChatGPT

June 16, 2023 Coin Telegraph

Meta claims Voicebox is the first AI that can generalize text-to-speech tasks it wasn’t trained to accomplish and describes it as a “breakthrough.”

Meta AI recently unveiled a “breakthrough” text-to-speech (TTS) generator it claims produces results up to 20 times faster than state-of-the-art artificial intelligence models with comparable performance.

The new system, dubbed Voicebox, eschews traditional TTS architecture in favor of a model more akin to OpenAI’s ChatGPT or Google’s Bard.

Among the main differences between Voicebox and similar TTS models, such as ElevenLabs Prime Voice AI, is that Meta’s offering can generalize through in-context learning.

Much like ChatGPT or other transformer models, Voicebox uses large-scale training datasets. Previous efforts to use massive troves of audio data have resulted in severely degraded audio outputs. For this reason, most TTS systems use small, highly-curated, labelled datasets.

Meta overcomes this limitation through a novel training scheme that ditches labels and curation for an architecture capable of “in-filling” audio information.

As Meta AI put in a June 16 blog post, Voicebox is the “first model that can generalize to speech-generation tasks it was not specifically trained to accomplish with state-of-the-art performance.”

This makes it possible for Voicebox to translate text to speech, remove unwanted noise by synthesizing replacement speech, and even apply a speaker’s voice to different language outputs.

According to an accompanying research paper published by Meta, its pre-trained Voicebox system can accomplish all of this using only the desired output text and a three-second audio clip.

The arrival of robust speech-generation comes at particular sensitive time as social media companies continue to struggle with moderation and, in the U.S., a looming presidential election threatens to once again test the limits of online misinformation detection.

Former U.S. president Donald Trump, for example, currently faces allegations that he mishandled confidential government materials after leaving office. Among the purported evidence cited in the case against him are audio recordings wherein he allegedly admitted to potential wrongdoing.

While there’s currently no indication that the former president intends to deny the content described in the audio files, his case illustrates that data integrity resides at the core of the U.S. legal system and, by extension, its democracy.

Voicebox isn’t the first tool of its kind, but it appears to be among the most robust. As such, Meta’s developed a tool for determining if speech was generated by it which the company claims can “trivially detect” the difference between real and fake audio. Per the blog post:

“As with other powerful new AI innovations, we recognize that this technology brings the potential for misuse and unintended harm. In our paper, we detail how we built a highly effective classifier that can distinguish between authentic speech and audio generated with Voicebox to mitigate these possible future risks.”

In the cryptocurrency world, AI has become as integral to day-to-day operations for most businesses as the internet or electricity. The largest exchanges rely on AI chatbots for customer interactions and sentiment analysis, and trading bots have become commonplace.

The advent of robust text-to-speech systems such as Voicebox, combined with automated trading, could help bridge a gap for would-be cryptocurrency traders who rely on TTS systems that, currently, may struggle with crypto jargon or multi-lingual support.

AI startup by ex-Meta and Google researchers raises $113M in seed funding

June 14, 2023 Coin Telegraph

AI startup by ex-Meta and Google researchers raises 3M in seed funding

The company is on a hiring spree and on the lookout for researchers, software engineers and product developers in AI.

A brand new artificial intelligence (AI) startup dedicated to rival ChatGPT creator, OpenAI, raised $113 million in seed funding, bringing up its valuation to $260 million within two months of inception.

Former AI researchers — previously working for Google DeepMind and Meta — co-founded Mistral AI, in May 2023 to develop open-source generative AI models. Arthur Mensch, the co-founder and CEO of the company said that the first round of funding “will give us the resources and network we need to start rolling out a new model of generative artificial intelligence.”

Prior to co-founding Mistral AI, Mensch was a research scientist at Google Deep Mind. The other two co-founders — Timothee Lacroix and Guillaume Lample — worked at Facebook AI as a research engineer and research scientist respectively.

*Mistral AI co-founders Guillaume Lample, Arthur Mensch, Timothée Lacroix (left to right). Source: Medium*

The funding round was led by Lightspeed Venture Partners and saw participation from JCDecaux Holding, Rodolphe Saadé and Motier Ventures among others. The trio will run the company from Paris and plan to release its first models for text-based generative AI in 2024.

*Mistral AI unfinished official website. Source: Mistra.AI*

The company is on a hiring spree and on the lookout for researchers, software engineers and product developers in AI. At the time of writing, the newly formed Mistral AI did not have any social media presence either.

Recently, Sam Altman, the CEO of OpenAI, met South Korean South Korean President Yoon Suk Yeol and urged the country to lead the manufacturing of chips dedicated to AI technology.

*OpenAI CEO Sam Altman and Korean President Yoon Suk Yeol shake hands at the Yongsan District, central Seoul presidential office on June 9.*

OpenAI currently utilizes chips from Taiwan, but Altman revealed the incoming need for a supply of chips from Korea in the future.

Magazine: Peter McCormack’s Real Bedford Football Club puts Bitcoin on the map

Quantum miners would yield ‘massive’ energy savings for blockchain: Study

June 8, 2023 Coin Telegraph

Bitcoin mining

University of Kent researchers compared three quantum systems to an ASIC miner, and the quantum machines were demonstrably more energy efficient.

A pair of scientists from the University of Kent’s School of Computing in the United Kingdom recently conducted a study comparing energy consumption rates for current ASIC-based miners to proposed quantum-based solutions.

According to the team’s preprint research paper, the systems utilizing quantum computing demonstrably outperformed standard mining rigs in energy efficiency:

“We show that the transition to quantum-based mining could incur an energy saving — by relatively conservative estimates — of about roughly 126.7 TWH, or put differently the total energy consumption of Sweden in 2020.”

Bitcoin mining operations alone consumed more than 150 terawatt hours annually (as of May 2022), per the paper, putting into perspective the potential impact the proposed quantum-based systems could have.

The pair’s conclusions were based on experiments comparing three different quantum mining systems to an Antminer S19 XP ASIC miner.

The quantum mining devices were split between a system featuring a single layer of fault tolerance, another one with two layers of fault tolerance and one without any dedicated error-correction features.

As the researchers point out, blockchain mining is one of the few areas of quantum computing where error correction isn’t such a big deal. In most quantum functions, errors create noise that functionally limit a computing system’s ability to produce accurate computations.

In blockchain mining, however, success rates with state-of-the-art classical systems are still relatively low. Per the research paper, “A classical Bitcoin miner is profitable with only a success-rate of about 0.000070%.”

The researchers also note that, unlike classical systems, quantum-based systems can actually be fine-tuned over time for increased accuracy and efficiency.

While quantum computing technology is still considered to be in its infancy, the very specific problem of blockchain mining doesn’t require a full-service quantum computing solution. As the researchers put it, “a quantum miner is not, and need not be, a scalable, universal quantum computer. A quantum miner need only perform a single task.”

Ultimately, the researchers conclude that it should be possible to build miners using existing quantum technologies that demonstrate quantum advantage over classical computers.

Despite the potential energy savings, it bears mention that the researchers focused on a type of quantum computing system called a “noisy intermediate-scale quantum” (NISQ) system.

According to the preprint paper, quantum miners should demonstrate “massive” energy savings at a size of around 512 quantum bits, or “qubits” — a term somewhat analogous to classical computing bits.

Typically, however, NISQ systems only operate with about 50-100 qubits, though there doesn’t appear to be an industry standard.

While the energy savings might be feasible, the costs of building and maintaining a quantum computing system in the 512 qubit range have, traditionally, been prohibitive for most organizations.

Only D-Wave and IBM offer client-facing systems in the same range (D-Wave’s D2 is a 512-qubit processor, and IBM’s Osprey weighs in at 433), but their architectures differ so greatly that comparisons between their qubit counts are ostensibly meaningless.

Fines and regulation: The ever-growing landscape of crypto compliance

June 5, 2023 Coin Telegraph

How have the regulatory and legal frameworks governing cryptocurrency fines transformed over the years, and what are the key trends and challenges?

It is a goal of many in the crypto industry to evolve the space and bring it into the mainstream, but the industry still faces constant criticism from and continues to be regulated by individuals who may not fully understand how it works. Regulatory bodies worldwide have been increasingly vigilant in addressing potential risks and ensuring compliance within the crypto space.

One notable aspect of this regulatory focus is the imposition of fines and penalties on individuals and companies involved in various crypto-related activities. Here’s how the landscape has transformed since 2018, leading to increasing regulation in 2023.

Increasing regulatory scrutiny

Since 2018, there has been a substantial rise in the number of regulatory actions and fines imposed on entities operating in the crypto sector, with the Securities and Exchange Commission in the United States and the Financial Conduct Authority in the United Kingdom strengthening their enforcement efforts with the aim to protect investors and maintain market integrity.

The expansion of regulatory frameworks, particularly those aimed at cryptocurrencies, has been a prominent development in 2023. They often include provisions related to Anti-Money Laundering (AML) and Know Your Customer (KYC) requirements, investor protection, and disclosure obligations. Implementing these regulations has enabled authorities to take a stronger stance against noncompliant individuals and companies.

The shift toward an individual-focused approach with a strong interest in the crypto sector has played a crucial role in driving the increase in enforcement actions and fines. By imposing fines on fraudulent activities, scams and unlawful touting, regulators aim to create a safer investment environment and deter bad actors from operating within the industry.

Notable cases and trends

Throughout this period, several high-profile cases have emerged involving celebrities, influencers and companies promoting cryptocurrencies without proper disclosures or engaging in fraudulent activities — such as those involving Floyd Mayweather Jr., DJ Khaled, Paul Pierce and Kim Kardashian. These cases have demonstrated the consequences of misleading the public regarding endorsements in the crypto space, serving as a wake-up call for both regulators and investors and highlighting the importance of transparency and due diligence.

Explore the Crypto Fines Database by Cointelegraph Research

Additionally, the rise of initial coin offerings a few years ago led to a surge in regulatory actions targeting projects that failed to comply with securities laws. Many ICOs were deemed unregistered securities offerings, resulting in fines, penalties, and even the shutdown of specific projects.

Implications

The landscape of crypto fines has changed substantially since 2018, reflecting the industry’s growing maturity and increasing regulatory scrutiny. Heightened enforcement efforts, expanding regulatory frameworks, and a focus on crypto users signify a shift toward a more regulated, responsible crypto ecosystem.

Regulations are continually tightening and evolving rapidly, with an increasing emphasis on AML/KYC compliance; hence, it’s vital to stay aware and navigate this changing landscape effectively. Cointelegraph Research’s comprehensive Crypto Fines Database is available to assist in ensuring compliance and avoiding potential fines by providing insights into the evolving regulatory environment. Learning from the past and staying proactive helps foster a more secure, trustworthy crypto ecosystem for all participants.

Scientists propose quantum proof-of-work consensus for blockchain

June 1, 2023 Coin Telegraph

Coin Telegraph

Boson sampling was once considered a problem looking for a solution. Now, it might be the bridge that brings quantum computing to the blockchain.

A team of researchers from universities in Australia and the United States, working in collaboration with quantum technology company BTQ, recently published research proposing a novel proof-of-work (PoW) scheme for blockchain consensus that relies on quantum computing techniques to validate consensus.

Our new paper on using NISQ-era quantum sampling problems in proof-of-work blockchain consensus algorithms. In collaboration with @BTQ_Tech.https://t.co/MKAB2czqSk
— Peter Rohde (@drpeterrohde) June 1, 2023

Dubbed “Proof-of-work consensus by quantum sampling,” the preprint research paper details a system that the authors claim “provides dramatic speedup and energy savings relative to computation by classical hardware."

According to the researchers, current algorithms for solving PoW consensus puzzles are slow and require a significant amount of computation resources to process:

“Whereas classical PoW schemes such as Bitcoin’s are notoriously energy inefficient, our boson sampling-based PoW scheme offers a far more energy efficient alternative when implemented on quantum hardware.”

According to the paper, the quantum advantage provided by this scheme would also increase the difficulty of mining, thus making it possible to “maintain consistent block mining time” as the number of miners increases, further incentivizing continuing participation of “quantum miners.”

The sampling process the researchers refer to, boson sampling, isn’t a new one, but its application to blockchain technology appears novel. Boson sampling has shown promise in numerous quantum computing applications. Still, as a non-universal quantum computing solution (it has to be used in a system built for a specific task), its potential has been limited to a select few domains, such as chemistry.

However, according to the researchers, it may be the perfect solution for future-proofing blockchain applications and, potentially, lowering the environmental impact of mining on the Bitcoin blockchain and similar chains.

Aside from quantum advantage, quantum hardware also has a leg up on old school computers due to the nature of how blockchain mining works.

One of the current advantages of classical supercomputers over their new quantum cousins is the ability to “precompute” when handling the same class of problem regularly. But, when it comes to blockchain, such precompute is essentially wasted.

Mining is, as the researchers put it, a problem that is “progress-free.” No matter how many times a blockchain puzzle is solved to provide proof-of-work, the computer and algorithms processing the challenges don’t ever get any better at solving the problem.

This means that quantum computers, despite being notoriously challenging to develop and expensive to build and maintain, would ultimately be capable of validating consensus more efficiently than state-of-the-art classical systems.

Researchers propose new scheme to help courts test deanonymized blockchain data

May 30, 2023 Coin Telegraph

Coin Telegraph

A team of researchers proposed five “argumentative schemes” designed to protect the rights of crypto crime suspects while also helping investigators.

A team of researchers from Friedrich-Alexander-Universität Erlangen-Nürnberg recently published a paper detailing methods investigators and courts can use to determine the validity of deanonymized data on the Bitcoin (BTC) blockchain.

The team’s preprint paper, “Argumentation Schemes for Blockchain Deanonymization,” lays out a blueprint for conducting, verifying and presenting investigations into crimes involving cryptocurrency transactions. While the paper focuses on the German and United States legal systems, the authors state that the findings should be generally applicable.

Bitcoin-related crime investigations revolve around the deanonymization of suspected criminals, a process made more challenging by blockchains’ pseudonymous nature. Users conducting blockchain transactions are identified by wallets (unique software addresses) instead of legal names.

However, blockchains are inherently transparent. Whenever data is added to a blockchain ledger, the transaction is recorded and made available for anyone with access to the blockchain to see.

Investigators trying to determine who is behind a specific wallet use the information ensconced in blockchain transactions (blocks) as data points that, when combined, form a digital paper trail.

According to the research team, the current bottleneck when it comes to these investigations is no longer a technological one; it’s a legal issue.

Law enforcement agencies have access to the tools needed to conduct preliminary blockchain analysis, but these early data points represent circumstantial evidence.

This evidence relies on certain raw assumptions that can only be validated by connecting on-chain activity to off-chain activity, such as compelling an exchange to disclose the identity or bank account information of users suspected of criminal involvement. Per the paper:

“In legal practice, those assumptions are critical for inferring the evidential value of the deanonymization of a perpetrator. However, no standard practice for deriving and discussing the reliability of those analysis results has been proposed yet.”

If conducted properly, blockchain investigations can reveal the perpetrator of a crime. The researchers cite the Wall Street Market case as an example. There, U.S. Postal Service investigators identified the operator of an illegal dark web marketplace by connecting various data points that law enforcement officers corroborated through surveillance operations.

However, the researchers state that such investigations risk impinging on suspects’ rights due to legal requirements. Prosecutors (in Germany and the U.S., per the paper) must demonstrate a certain degree of evidence of guilt before a warrant for invasive investigations, such as surveillance or arrests, be issued.

To aid investigators and prosecutors while also ensuring the law is applied fairly to suspects, the researchers propose a standard framework containing five argumentative schemes designed to ensure proper reporting and explanation throughout the legal process.

*Two of the schemes explored by researchers. Source: “Argumentation Schemes for Blockchain Deanonymization"*

The above image shows two of the schemes, each utilizing a set of defined premises to frame a specific conclusion and then providing a set of critical questions to assess the strength of the argument.

The researchers assert that “by utilising the schemes, an analyst can clearly articulate the employed heuristics, their individual strengths, and potential weaknesses. This increases the comprehensibility of such analyses and court proceedings for the decision makers, and also eases the documentation for later verification by an expert witness.”

Neuralink gets FDA approval for ‘in-human’ trials of its brain-computer interface

May 26, 2023 Coin Telegraph

The Neuralink device is designed to be surgically implanted and, according to CEO Elon Musk, will eventually be marketed to the general public.

Elon Musk’s brain-computer interface (BCI) startup, Neuralink, reportedly received approval from the United States Food and Drug Administration (FDA) to conduct some form of “in-human” trials for its surgically implanted brain device.

Congratulations Neuralink team! https://t.co/AWZGf33UDr
— Elon Musk (@elonmusk) May 26, 2023

It’s unclear at this time what kind of trials the Neuralink tweet is referring to — the only other information mentioned was that the company isn’t accepting applicants yet.

Neuralink was denied the FDA’s approval for human trials in early 2022 over several concerns related to safety, which, at the time, the company was reportedly working to resolve.

The FDA’s apparent U-turn would indicate that such concerns have been worked out, and the company should then be free to conduct limited testing of its surgically implanted BCI device.

In describing the function of the BCI, Elon Musk has stated it would help with certain medical and mobility issues. BCIs are developed for the treatment of numerous conditions, such as epilepsy. They also provide quality-of-life services for disabled persons, such as the ability to direct a mouse cursor via eye movement or thought.

Musk has also said the device would be developed for use by the general public as a general-purpose BCI, which allows humans to interface with machines via thought and protect us from being replaced by machines.

Until now, Neuralink has only been allowed to conduct testing on laboratory animals such as monkeys and pigs. The company recently faced a federal probe over alleged animal mistreatment; however, Neuralink denied any wrongdoing, and the investigation appears to have ended quietly. A separate investigation over the alleged transportation of contaminated devices remains ongoing.

With human testing on the table, Musk’s vision for hybrids of humans and artificial intelligence (AI) comes one step closer. It therefore might be time for developers and entrepreneurs to start considering the applications and use cases of a recreational, surgically implanted BCI.

Musk has also proposed that BCIs would allow people to operate smartphones faster with their minds than they could with their thumbs. Recent research from the University of Texas has demonstrated that AI systems similar to ChatGPT can already be trained to interpret brain waves to a limited degree.

But perhaps the most interesting potential use case for the cryptocurrency and blockchain communities comes from Microsoft. Several patents filed in 2018 and 2019 describe a system by which a wearable “sensor” could be engineered to provide “proof-of-work” for cryptocurrency rewards and blockchain mining and validation.

According to one of the patent filings:

“For example, a brain wave or body heat emitted from the user when the user performs the task provided by an information or service provider, such as viewing advertisement or using certain internet services, can be used in the mining process.”

With a BCI like the one Neuralink describes in its original research paper and a paradigm like the one described in the Microsoft patent, it should be possible to validate brain waves natively, thus making it possible to verify “proof-of-work” via thought alone.

Crypto remains hopeful as market moves sideways: Report

May 8, 2023 Coin Telegraph

Coin Telegraph

BTC continues to trade in a narrow range, and memecoins steal the limelight from NFTs as views on the macroeconomic outlook diverge.

After a turbulent month for the crypto industry in March, Bitcoin’s (BTC) price went sideways in April despite some volatility. The meteoric rise of memecoins, such as PEPE, made headlines, and First Republic, another mid-sized United States bank, went under. However, on the basis of current market sentiment is a standoff between markets and policymakers: While the U.S. Securities and Exchange Commission Chairman Jerome Powell publicly states that interest rates are unlikely to come down this year, the markets for risk-on assets like crypto have firmly priced in a pivot in the coming months.

In times like these, it is wise to drill deeper into the fundamentals that will shape future market movements. With an uncertain macro environment and a looming regulatory crackdown in the U.S., there are other notable developments that are easily drowned out by these dominant news items.

The report is available for free on the Cointelegraph Research Terminal.

For those keen to gain a deeper understanding of the crypto space’s various sectors, Cointelegraph Research publishes a monthly Investors Insights Report that dives into venture capital, derivatives, decentralized finance (DeFi), regulation and much more. Compiled by leading experts on these various topics, the monthly reports are an invaluable tool to quickly get a sense of the current state of the blockchain industry.

NFT hype fades as memecoins take over

Nonfungible token (NFT) collectibles are one of the few sectors that took a major hit this month. Memecoins, such as PEPE, may be partially responsible for this, as they absorbed the attention, printing eye-watering gains. BRC-20 tokens, a new abstraction created on the Bitcoin Ordinals protocol, may also compete for cash inflow from traditional NFT collectibles traders. Sellers have started to persistently outnumber buyers on NFT marketplaces recently, and this trend is likely to continue.

There are concerns about the NFT market going into free fall, as all important metrics, such as volume and active wallets, have been on a steep decline. NonFungible reported only 49,200 active wallets and a sales volume of $80,500 this month. The NFT marketplace wars, combined with diminishing excitement around NFTs, are other driving factors behind this long-term development.

Despite the overall NFT market slump, a niche NFT sector that is picking up steam is the NFT lending market. Since the start of 2022, this sector has witnessed double-digit growth every month, and this continued in April with a 16.13% increase in new users.

Mining stocks outperform BTC

Every Cointelegraph Research Monthly Trends Report includes coverage of mining economics and crypto stocks. For investors interested in increasing their exposure to BTC, mining stocks have historically been a popular option. While idiosyncratic factors have negatively impacted individual stocks this month, the sector as a whole seems to have exited from the 2022 bear market.

The highest returns were again recorded by TeraWulf, which continued its rally with another 85% rise in evaluation. CleanSpark, IrisEnergy and BitDigital were other strong gainers. Notably, the stocks in April outperformed BTC on aggregate after lagging behind in the previous month. Where Bitcoin only posted a 2.8% close, the largest crypto stocks, dominated by mining, recorded 12.9%

Of course, increased evaluations in the mining industry are highly sensitive to BTC’s price action. For those with confidence in improving macroeconomic conditions for risk-on assets, these stocks may offer good entries as they were previously battered by the bear market. The stocks section of the monthly report tracks the fundamentals of major companies in the industry and thus amends our regular analysis of Bitcoin mining economics.

The Cointelegraph Research team

Cointelegraph’s Research department comprises some of the best talents in the blockchain industry. Bringing together academic rigor and filtered through practical, hard-won experience, the researchers on the team are committed to bringing the most accurate, insightful content available on the market.

Demelza Hays, Ph.D., is the director of research at Cointelegraph. Hays has compiled a team of subject matter experts from finance, economics and technology to bring the premier source for industry reports and insightful analysis to the market. The team utilizes APIs from various sources to provide accurate, useful information and analyses.

With decades of combined experience in traditional finance, business, engineering, technology and research, the Cointelegraph Research team is perfectly positioned to put its combined talents to proper use with the latest Investor Insights Report.

Tim Cook says Apple will weave AI into products as researchers work on solving bias

May 5, 2023 Coin Telegraph

The Apple CEO said the company would “continue weaving” AI into its products as internal research indicates an emphasis on building unbiased AI systems.

CEO Tim Cook gave a rare, if guarded, glimpse into Apple’s walled garden during the Q&A portion of a recent earnings call when asked his thoughts on generative artificial intelligence (AI) and where he “sees it going.”

Cook refrained from revealing Apple’s plans, stating upfront, “We don’t comment on product roadmaps.” However, he did intimate that the company was interested in the space:

“I do think it’s very important to be deliberate and thoughtful in how you approach these things. And there’s a number of issues that need to be sorted. … But the potential is certainly very interesting.”

The CEO later added the company views “AI as huge” and would “continue weaving it in our products on a very thoughtful basis.”

Cook’s comments on taking a “deliberate and thoughtful” approach could explain the company’s absence in the generative AI space. However, there are some indications that Apple is conducting its own research into related models.

A research paper scheduled to be published at the Interaction Design and Children conference this June details a novel system for combating bias in the development of machine learning datasets.

Bias — the tendency for an AI model to make unfair or inaccurate predictions based on incorrect or incomplete data — is oft-cited as one of the most pressing concerns for the safe and ethical development of generative AI models.

So glad OpenAI is keeping its bias in check. pic.twitter.com/y4a7FUochR
— Brooklin Nash (@realBrookNash) April 27, 2023

The paper, which can currently be read in preprint, details a system by which multiple users would contribute to developing an AI system’s dataset with equal input.

Status quo generative AI development doesn’t add in human feedback until later stages, when models have typically already gained training bias.

The new Apple research integrates human feedback at the very early stages of model development in order to essentially democratize the data selection process. The result, according to the researchers, is a system that employs a “hands-on, collaborative approach to introducing strategies for creating balanced datasets.”

It bears mention that this research study was designed as an educational paradigm to encourage novice interest in machine learning development.

It could prove difficult to scale the techniques described in the paper for use in training large language models (LLMs) such as ChatGPT and Google Bard. However, the research demonstrates an alternative approach to combating bias.

Ultimately, the creation of an LLM without unwanted bias could represent a landmark moment on the path to developing human-level AI systems.

Such systems stand to disrupt every aspect of the technology sector, especially the worlds of fintech, cryptocurrency trading and blockchain. Unbiased stock and crypto trading bots capable of human-level reasoning, for example, could shake up the global financial market by democratizing high-level trading knowledge.

Furthermore, demonstrating an unbiased LLM could go a long way toward satisfying government safety and ethical concerns for the generative AI industry.

This is especially noteworthy for Apple, as any generative AI product it develops or chooses to support would stand to benefit from the iPhone’s integrated AI chipset and its 1.5 billion user footprint.

Research

Blockchain infrastructure is still on top

Increasing regulatory scrutiny

Notable cases and trends

Implications

NFT hype fades as memecoins take over

Mining stocks outperform BTC

The Cointelegraph Research team

Share this video