OpenAI has announced the ability to fine-tune its powerful language models, including both GPT-3.5 Turbo and GPT-4.
The fine-tuning allows developers to tailor the models to their specific use cases and deploy these custom models at scale. This move aims to bridge the gap between AI capabilities and real-world applications, heralding a new era of highly-specialised AI interactions.
With early tests yielding impressive results, a fine-tuned version of GPT-3.5 Turbo has demonstrated the ability to not only match but even surpass the capabilities of the base GPT-4 for certain narrow tasks.
All data sent in and out of the fine-tuning API remains the property of the customer, ensuring that sensitive information remains secure and is not used to train other models.
The deployment of fine-tuning has garnered significant interest from developers and businesses. Since the introduction of GPT-3.5 Turbo, the demand for customising models to create unique user experiences has been on the rise.
Fine-tuning opens up a realm of possibilities across various use cases, including:
Improved steerability: Developers can now fine-tune models to follow instructions more accurately. For instance, a business wanting consistent responses in a particular language can ensure that the model always responds in that language.
Reliable output formatting: Consistent formatting of AI-generated responses is crucial, especially for applications like code completion or composing API calls. Fine-tuning improves the model’s ability to generate properly formatted responses, enhancing the user experience.
Custom tone: Fine-tuning allows businesses to refine the tone of the model’s output to align with their brand’s voice. This ensures a consistent and on-brand communication style.
One significant advantage of fine-tuned GPT-3.5 Turbo is its extended token handling capacity. With the ability to handle 4k tokens – twice the capacity of previous fine-tuned models – developers can streamline their prompt sizes, leading to faster API calls and cost savings.
To achieve optimal results, fine-tuning can be combined with techniques such as prompt engineering, information retrieval, and function calling. OpenAI also plans to introduce support for fine-tuning with function calling and gpt-3.5-turbo-16k in the upcoming months.
The fine-tuning process involves several steps, including data preparation, file upload, creating a fine-tuning job, and using the fine-tuned model in production. OpenAI is working on a user interface to simplify the management of fine-tuning tasks.
The pricing structure for fine-tuning comprises two components: the initial training cost and usage costs.
The introduction of updated GPT-3 models – babbage-002 and davinci-002 – has also been announced, providing replacements for existing models and enabling fine-tuning for further customisation.
These latest announcements underscore OpenAI’s dedication to creating AI solutions that can be tailored to meet the unique needs of businesses and developers.
(Image Credit: Claudia from Pixabay)
See also: ChatGPT’s political bias highlighted in study
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with Digital Transformation Week.
Explore other upcoming enterprise technology events and webinars powered by TechForge here. Tags: ai, artificial intelligence, fine tuning, gpt-3, gpt-3.5 turbo, gpt-4, openai
Key Points Google said the chatbot now includes 40 languages and will include audio responses. The improvements come as Google tries to stay ahead of rapidly improving competition from Microsoft-backed OpenAI, Anthropic, and others in the white-hot generative AI space.
A hand holding a smartphone with the Google Bard logo is seen, in front of a computer screen showing Google’s logo.Nikos Pekiaridis | Nurphoto | Getty Images
Google is adding new features to its chatbot Bard as it tries to stay ahead of rapidly improving competition from Microsoft-backed OpenAI, Anthropic, and others in the white-hot generative AI space. The search giant debuted its ChatGPT competitor to the public in March, starting with tests in the U.S. and the U.K. Since then, the company announced new updates at its annual developer conference Google I/O in May. OpenAI’s ChatGPT also announced new features in May, including launching the ChatGPT app for iOS and integrating voice prompts. Here are a few of the updates Google launched Thursday: Analysis of images: Bard will now have image capabilities, which the company announced as a Google Lens feature at Google IO in May. Users will be able to upload an image and ask Bard for information about an image or ask it to make a caption based on the image, the company’s blog post states.The feature is only available in the English language.
Different styles of responses: The company says Bard users can now change the tone and style of Bard’s responses to five different options: simple, long, short, professional or casual. The company gives the example of asking Bard to help you write a marketplace listing for a vintage armchair, and then shorten the response. This feature is live in English and will expand to new languages soon, it says.
Read responses: Users can use audio to listen to Bard’s responses, which the company says is helpful for users who want to hear the correct pronunciation of a word or listen to a poem or script. This feature works in over 40 languages, the company stated.
More languages and regions: The company said people can use Bard in over 40 languages, including Arabic, Chinese, German, Hindi and Spanish. It’s also available now in Europe, which was reportedly delayed due to The Irish Data Protection Commission voicing privacy concerns. “We’ve proactively engaged with experts, policymakers and privacy regulators on this expansion,” the company said in its Thursday blog post.
OpenAI is an artificial intelligence technology company founded in 2015 with the mission of developing advanced AI technologies to bring more value and convenience to humanity. OpenAI’s technologies are based on techniques such as deep learning and natural language processing, and are trained and optimized with vast amounts of data to achieve powerful AI applications.
OpenAI’s technology is widely applied in various fields, including natural language processing, computer vision, reinforcement learning, and robotics. For example, in the field of natural language processing, OpenAI has developed a series of powerful language models that can perform tasks such as text generation, question-answering systems, and conversation. In computer vision, OpenAI has developed advanced visual models that can achieve image recognition, image generation, and other applications. In reinforcement learning, OpenAI has developed advanced reinforcement learning algorithms that can be applied in intelligent robots and automatic control systems.
The strength of OpenAI’s technology lies in its powerful capabilities and broad range of applications. OpenAI’s technology can automatically identify and process various language and image features, extract useful information from them, and generate high-quality text and images. OpenAI’s technology can be applied in various fields, including science, engineering, medicine, finance, and business, to bring more value and convenience to various industries.
Although OpenAI’s technology has achieved great success, there are still some limitations. Because OpenAI’s technology is based on training with vast amounts of data, it may not provide precise answers when faced with new problems. In addition, OpenAI’s technology also poses some security risks and ethical issues that require further research and exploration.
In conclusion, OpenAI is a highly influential artificial intelligence technology company that has achieved broad application in various fields. OpenAI’s technology can bring more value and convenience to people, and its limitations, such as potential data biases and ethical concerns, are still being researched and addressed.
The cryptocurrency market is in turmoil, exacerbated by the collapse of luna and the UST stablecoin, both tied to the terra blockchain.
Why it matters
Over $17 billion in crypto value has been wiped out, and the collapse has raised questions about stablecoins in general. The crash has caught the eye of politicians and regulators.
What’s next
The company behind UST will be building a new blockchain, though it won’t include a stablecoin.
The cryptocurrency market is brutal right now: Look in any direction and you’ll see red charts. Bitcoin has ended up in the red for eight consecutive weeks, a record for the cryptocurrency, and ether is at its lower point since 2020. While it’s painful for crypto investors, this dive isn’t entirely unprecedented. Cryptocurrencies are infamous for their volatility, and tempestuous economic conditions are bringing down not just crypto, but the stock market too.
What is unprecedented, however, is the collapse of the luna cryptocurrency and its associated terraUSD stablecoin, aka UST. You may not have heard of UST before, or know what a stablecoin is, but it’s a big deal. Billions of dollars in crypto wealth have been vaporized, sending shockwaves throughout the whole market.
There are two intertwined stories here: That of the UST stablecoin and that of luna, both of which are part of the Terra blockchain. The UST coin is designed to retain a value of $1 at all times, but it was depegged on May 9, and has since fallen to just 7 cents. Then there’s luna, the centerpiece of Terra’s ecosystem. Its value has collapsed in one of the most stunning crypto crashes ever recorded.
Luna’s price chart depicts a historic crash. Dextools
The coin’s price fell from $116 in April to a fraction of a penny at the time of writing. Such an implosion has been seen in small-cap memecoins in the past, but never for something the size of luna, which had a market cap of over $40 billion just last month.
“This is historic for the crypto markets,” said Mike Boroughs, co-founder of crypto investments firm Fortis Digital. “This is a defining moment for the space due to its size and impact in terms of the amount of people that lost substantial value.”
The Terra saga has bigger implications than the lost billions. It’s brought up questions about similar tokens namely Tether, and regulators across the political isle have their eyes set on stablecoins. Here’s what you need to know.
What’s a stablecoin?
To understand the crypto catastrophe, you first need to know what a stablecoin is. In essence, it’s a cryptocurrency that’s pegged to a more stable currency. The biggest such coins are tether and USDC, which like most stablecoins are both tied to the US dollar. So if you have 1,000 USDC tokens, for instance, they can at any time be exchanged for $1,000.
Stablecoins are integral parts of “DeFi,” or decentralized finance, designed to be ways for investors to hedge against the volatility of the cryptocurrency market. Say ether’s price is $2,000 — a trader could exchange one ether for 2,000 USDC tokens. If tomorrow ether drops 50% to $1,000, those 2,000 USDC tokens would still be worth $2,000 and could be traded for two ether tokens. When investors smell a downswing coming, they put their money on stablecoins like tether, USDC and, until this week, UST.
Stablecoins also provide the means for cryptocurrency loaning and borrowing, making them a foundational technology of DeFi.
The UST coin, created by Terraform Labs, is different from tether and USDC in a key way — it’s not backed by actual US dollars, but rather is what’s known as an algorithmic or decentralized stablecoin. (Tether’s US reserves have come under scrutiny in the past, with there being some conjecture over how many dollars it actually holds — but it’s US dollar backed in principal.) For Terraform Labs, the idea was that through a few clever mechanisms, plus billions in bitcoin reserves, the UST’s dollar peg could be maintained without it having to be backed by the dollar.
“A decentralized stablecoin is the Holy Grail of DeFi,” said Cyrus Younessi, former head of risk management at MakerDAO, the group behind DAI stablecoin. The selling point of bitcoin and ether is that they’re difficult for bureaucrats, politicians and central bankers to control, but their downside is price volatility. “If you could take those assets, extract stability out of them and productize it, then that’s huge,” Younessi said.
“But it’s not very viable.”
Terra, luna and UST: What are they?
Terra is a blockchain, just like ethereum and bitcoin. While ethereum’s blockchain natively produces ether tokens, terra natively produces luna. In the deys preceeding the depeg, luna was trading at $85.
To create UST, you need to burn luna. So for instance, in early May you could trade one luna token for 85 UST (since luna was worth $85), but the luna would be destroyed (“burned”) in the process. This deflationary protocol was meant to ensure luna’s long-term growth. As more people buy into UST, more luna would be burned, making the remaining luna supply more valuable.
To entice traders to burn luna to create UST, creators offered an insane 19.5% yield on staking — which is essentially crypto terminology for earning 19.5% interest on a loan — through what they called the Anchor Protocol. Instead of parking your savings at a bank for a 0.06% interest rate, the pitch is to turn put your money into UST, where it can earn nearly 20% in interest. Before the depegging, over 70% of UST’s circulating supply, around $14 billion, was deposited in this scheme.
Here’s the key to UST retaining its peg: 1 UST could always be exchanged for $1 worth of luna. So if UST slipped to 99 cents, traders could profit by buying a huge amount of UST and exchanging it for luna, profiting 1 cent per token. The effect works in two ways: People buying UST drives the price up, and UST being burned during its exchange to luna deflates the supply.
Then there’s the reserves. Terraform Labs founder and CEO Do Kwon created the Luna Foundation Guard, a consortium whose job it is to protect the peg. The LFG had about $2.3 billion in bitcoin reserves, with plans to expand that to $10 billion worth of bitcoin and other crypto assets. If UST dipped below $1, bitcoin reserves would be sold and UST bought with the proceeds. If UST goes above $1, creators would sell UST until it goes back to $1, with the profit being used to buy more bitcoin to pad out the reserves.
It all makes sense. But UST, at the time of writing, is worth 7 cents. How?
The depeg of UST
It all started on Saturday, May 7. Over $2 billion worth of UST was unstaked (taken out of the Anchor Protocol), and hundreds of millions of that was immediately sold. Whether this was a reaction to a volatile period — the rise in interest rates has particularly affected cryptocurrency prices — or a more malicious attack on Terra’s system is a topic of debate.
Such huge sells pushed the price down to 91 cents. Traders tried to take advantage of arbitrage, exchanging 90 cents worth of UST for $1 worth of luna, but then a speed bump appeared. Only $100 million worth of UST can be burned for luna per day.
Investors, already flighty in the current gloomy market, flocked to sell their UST once the stablecoin couldn’t retain its peg. It bounced between 30 cents and 50 cents in the week following the initial depeg, but has now fallen to a steady low of under 20 cents. Its market cap, which was around $18 billion in early May, now stands at $770 million.
It’s worse for luna holders. The value of luna tokens has almost completely disappeared: After reaching a high of just under $120 in April, luna’s current price is less than a fiftieth of a penny.
On the possibility of this being a malicious attack. Some have speculated that an attacker attempted to break UST in order to profit from shorting bitcoin — that is, betting on its price going down. If would-be attackers created a large position in UST and then unstaked $2 billion at once, it could depeg UST, which would mean terra’s team would have to sell portions of its bitcoin reserve to repeg the stablecoin. Once investors saw that UST lost its peg, they would then rush to unstake and sell their UST, which would require more bitcoin reserves to be sold, adding further sell pressure.
Again, this is still speculation. Younessi is unsure whether the depeg was caused by a coordinated attack or not, but said that the responsibility is on crypto developers to create more secure systems.
“Our job as DeFi builders is to build systems that are resistant [to exploits],” he said. “That’s literally in the original threat model that anyone in crypto builds: How would this hold up if a guy with $100 billion came in and tried to take this down?”
“Terra could have grown to be 10 times as large” before such a crash, he said to CNET. “Better that we prick that bubble of unsustainable protocols sooner than later.”
Why does it matter?
This matters for three reasons.
First, over $17 billion in crypto value has been wiped out through luna and UST alone. There have been anecdotal reports of self-harm by those who had most of their savings staked in UST — though these can’t be confirmed, it’s clear that a lot of people lost a lot of money in the collapse. The damage isn’t contained to Terra’s ecosystem though, as Fortis Digital’s Boroughs notes. Many who were exposed to luna and UST would have sold off big parts of their crypto portfolio to recoup some of the damage, pulling the entire market down.
Second, it raises questions about other stablecoins. Again, UST was unusual in that it was an algorithmic stablecoin, unlike tether and USDC. But the stability of those coins has always been somewhat in doubt: For instance New York’s attorney general last year accused tether, the biggest stablecoin, of lying about how much it actually held in dollar reserves. An estimated $10 billion of tether has been withdrawn since May 11, leading some to speculate that it may be the second stablecoin to depeg.
Boroughs worries that, if UST was attacked, similar plays could be made against the others.
“The question in our minds becomes, does what happened to UST spread to other stablecoins?” he said. “If big whales found a playbook here that works to attack UST, we worry they may reuse that playbook in other areas of the market.”
Last, and possibly most significantly, the collapse of UST has caught the attention of powerful politicians and regulators. Secretary of the Treasury Janet Yellen said on May 10 that UST’s depegging “simply illustrates that this [stablecoins] is a rapidly growing product and there are rapidly growing risks.”
The desire to regulate stablecoins appears to have united the US’ two parties: Bloomberg reports that Wyoming Republican Cynthia Lummis and New York Democrat Kirsten Gillibrand are preparing a bipartistan proposal that would seek to stop a blow up like the one suffered by UST holders from happening again.
What’s next for terra, luna and UST?
It’s been a rough week for terra developers since UST depegged. After some, including Binance CEO Changpeng Zhao, questioned how the project’s bitcoin reserves were used, the Luna Foundation Guard, the consortium setup to protect UST’s peg, tweeted that its bitcoin reserves fell from 80,000 (about $2.2 billion) to just 313 ($9.2 million).
The remaining reserves will be used to “compensate remaining users of UST, smallest holders first.”
Meanwhile, Kwon, the Terraform Labs CEO, has controversial plan to revive luna.
Kwon proposed a plan to “fork” the terra blockchain. In essence, this means the creation of a new blockchain that’s modeled on the previous one, with some key changes. The new blockchain would be created alongside a billion luna tokens, which would be distributed among current luna and UST holders, as well as fund development of new terra apps.
Perhaps conceding the foundational problem of tethering luna to UST, Kwon proposed removing UST, previously the main selling point of the blockchain, from the terra ecosystem. “Terra’s app ecosystem contains hundreds of developers working on everything from DeFi to fungible labor markets, state-of-the-art infrastructure and community experience,” he said, proposing this should be preserved at the expense of terraUSD.
The most famous fork in crypto history happened to ethereum in 2016. After a hacker robbed 3.6 million ether from a DAO — then worth $50 million, now worth over $7 billion — ethereum’s developers forked the blockchain, creating a new chain identical in all ways except the restoration of the stolen million ether. It caused a rift within the community, with some maintaining the original chain to this day, calling it Ethereum Classic.
Kwon’s plan is a nod to that fork. The new plan will see the current Terra blockchain renamed to Terra Classic, while the new chain will simply go by Terra.
Many, including Zhao, are skeptical that the plan will work, but UST and luna holders voted in favor of the proposal. A new era for luna will begin around May 27, when Terraform Labs hopes to launch the new ecosystem.
An NFT is an abbreviation of Non-Fungible Token that encrypts an image. The NFT term in its abbreviation signifies that the image protected by it is non replaceable and unique. An NFT can be any digital image,including videos, or music, but a lot of the current excitement is around using the technical means to encrypt images to sell digital art and in most cases it is about digital images or short movies, GIFs etc.
At a very high level, most NFTs are part of the Ethereum blockchain. Ethereum is a cryptocurrency, like bitcoin or dogecoin, but its blockchain also supports these NFTs, which store extra information that makes them work differently from, say, an ETH coin. It is worth noting that other blockchains can implement their own versions of NFTs.
Christopher Torres, NYAN-CAT, 2011
The most famous NFT so far became the Dogecoin, which isn’t an NFT. But the GIF below of a dogecoin above is one such NFT and it became one of the most famous ones.
NFTs are designed to give the collectors something that can’t be copied: ownership of the work (though the artist can still retain the copyright and reproduction rights, just like with physical artwork). To put it in terms of physical art collecting: anyone can buy a Monet print. But only one person can own the original.
Recently, Mike Winkelmann, better known as Beeple, has sold the most expensive work of digital art in history.
It’s part of an explosion in the market for NFTs, or non-fungible tokens — digital tokens that prove ownership of things like the images sold by Beeple, that can’t even be touched and yet they are sold to collectors as unique original artworks.
“I honestly, like, I never thought I could sell my work,” says Beeple in an interview at his home in South Carolina in October 2020. Two months later, in December 2020, he reached a price of $3.5 million, selling digital art encrypted by an NFT.
Mike Winkelmann aka Beeple, Collage, 2021
In March 2021, Christie’s, a 225-year-old and somewhat very traditional auction house that previously only sold mostly physical art, auctioned an entirely digital piece by Beeple. It sold for $69,346,250.
“If everybody wants it, well, then it has value,” is the comment of Beeple for this sale, that shocked the entire art world.
The speculation in this market became so crazy, that when a $95,000 Banksy physical artwork was recently burned and turned into an NFT, the NFT was sold for nearly $400,000. A cat meme recently sold for $600,000. To understand who’s paying these prices, it’s important to understand NFTs.
Unlike Bitcoin, which are all identical by design, NFTs are unique. To some degree, what NFTs offer for sale is the idea of scarcity. It’s possible to buy a token that represents art in the physical world, but NFTs can also protect and encrypt digital assets like an image or a tweet.
“So in May 1, 2007, I started doing a sketch a day, every single day, start to finish, and uploading it online,”Winkelmann said. “And after a year of that, I learned a lot about drawing. Like, I got much better at drawing. I was still very, very bad, as you can see from the Christie’s piece. But I learned a lot.”
Mike Winkelmann aka Beeple, Everyday Crap, 2021
Who is in fact Beeple?
Beeple AKA Mike Winkelmann is a digital artist who has risen to fame from his “Everydays” project, where he creates a new piece of art every single day. Beeple has been making and posting one piece of art a day for over 13 years, as he began in 2007, which means his body of work contains more than 5000 images.
His project, Everydays, was billed at the auction house as “a unique work in the history of digital art.” The official title of the work is: “Everydays — The First 5000 Days.” As his fame grew online, Beeple has collaborated with Louis Vuitton and pop stars like Justin Bieber and Katy Perry.
Mike Winkelmann aka Beeple, Everydays, Miami, 2021
Figma Inc., the startup behind a cloud service used by Microsoft Corp. and Uber Technologies Inc. to design interfaces for their applications, landed a funding round of $ 200 million for a valuation of $ 10 billion.
Figma shared the news with Bloomberg today. According to the publication, the investment included participation from Durable Capital Partners, Morgan Stanley and Sequoia, who also participated in the startup’s previous $ 50 million fundraiser last April.
Between Figma’s stealth launch in December 2015 and its $ 50 million fundraising round last April, the startup has built an installed base of over 4 million users. Its clients include many of the biggest players in the tech industry. Figma’s platform is primarily used by corporate software design teams, which rely on it to create sketches of an interface and then turn them into an interactive prototype that works as if it were of an application.
Figma provides a number of unique features for the sketching and prototyping phase of a project. When creating the initial mockup of an interface, users have access to a technology called vector networks, which Figma presents as an improved version of the digital pen found in most graphic design programs. The startup’s vector networks simplify some common design tasks such as drawing geometric shapes.
For the prototyping phase of projects, Figma offers a feature called Auto Layout. The technology automatically transforms design details, such as the distance between two buttons or their location in the interface, into code, speeding up application projects by reducing the need for developers to perform the task manually. Figma’s platform also automates some of the tasks involved in updating an interface prototype. For example, if a designer drags a list item that was originally in the center of the list up, Figma can automatically rearrange the other entries.
Built on top of Figma’s core design capabilities is a collaborative feature set. Several designers can edit a file at the same time and exchange inputs via a built-in commenting system.
“Our vision is to make design accessible to everyone,” Figma co-founder and CEO Dylan Field wrote in a statement. blog post today. “It means creating expert tools for designers and also taking on all the roles involved in product development. But above all, it’s about opening access and empowering teams to think, feel and work in a design-driven way.
Figma plans to increase its workforce to 500 employees by the end of the year, according to Bloomberg, more than triple the number of employees it had at the start of 2020. The startup also hinted that it was considering to make acquisitions.
Figma has raised over $ 332 million in total funding to date. Its main competitors are Adobe Inc. and InVision Inc., a publicly traded startup that received a valuation of $ 1.9 billion after its last funding round in 2018.
A Nokia OYJ ultra deployable 5G Massive MIMO millimeter wave antenna. Photographer: Patrick T. Fallon/Bloomberg
The Finnish state is buying shares of Nokia Oyj to send a signal that it’s ready to protect the maker of 5G mobile networks amid a geopolitical battle in which the U.S. has expressed interest in owning a part of the company.
The state’s equity-asset manager Solidium Oy recently crossed the 5% threshold in Nokia shares, and now has about 1 billion euros ($1.2 billion) worth of stock. The telecommunications-equipment maker was earlier this year reported to have hired advisers to consider strategic alternatives and U.S. Attorney General William Barr has suggested his government should buy a stake in Nokia or its rival Ericsson AB.
Solidium’s stock purchases “act as a counter balance” to Barr, and “show William Barr and his European colleagues that the situation is under control,” Chief Executive Officer Antti Makinen said. “There’s a domestic anchor owner who can safeguard continuity, at least to some extent.”
Nokia, Sweden’s Ericsson and China’s Huawei Technologies Co. are locked in intense competition to supply 5G technology that’s increasingly been caught up in a global political battle. U.S. interest on Nokia and Ericsson centers on countering the threat posed by China’s dominance of 5G, where it has no domestic supplier. Other governments, including the U.K., have also banned Huawei from supplying 5G networks.
“From a European perspective, it’s important that Europe has 5G technology. But Europe isn’t an entity who could buy shares,” Makinen said. “For the Finnish state, it boils down to this: it’s one of the most — if not the most — important company in Finland and keeping it Finnish has value to us.”
Solidium won’t disclose plans of further purchases, and Makinen said the 5% it now has “feels like a suitable amount.”
The bet on Nokia comes with “plenty of risks,” both technological and trade-policy related, he said. “This isn’t a risk-free investment by any means.”
A trial of the English coronavirus app is getting under way.
It will be limited to residents in the Isle of Wight, the London Borough of Newham and NHS volunteer responders to begin with.
The app will be available in Apple and Google’s online stores, but users will need to enter a code to activate it.
The software will tell users to self-isolate for a fortnight if the app detects they have been close to someone else diagnosed with the virus.
Baroness Dido Harding – who heads up the wider Test and Trace initiative – had earlier voiced concern about implementing the automated contact-tracing feature because of fears many people who had been falsely flagged might be told to go into quarantine.
The app has several other functions, including:
An alert system that informs users of the coronavirus risk level close to their home, with the area defined by the first part of their postcode
A QR barcode scanner, so users can check in when they visit a venue and be told if others there later tested positive
A symptom-checking tool, which allows users to book a free test and get the results via the app
A countdown function that comes into effect if they are told to self-isolate, so users can keep track of how long to stay at home
It initially works in five languages, with plans to add more soon.
The contact-tracing element of the software is based on Google and Apple’s privacy-centric system.
The developers acknowledge there are still issues with measuring the distance between handsets, meaning some people will be incorrectly logged as being at high risk.
But when trying to detect this, lab tests indicate:
31% of cases are missed when the handsets were within range
45% of cases are incorrectly flagged when the two handsets were in fact further apart
However, if the boundary is set at 5m, the accuracy rates radically improve.
Then the handsets detect each other in more than 99% of all cases, regardless of whether iPhones or Android devices were involved.
This is not useful in practice, but indicates the flaw that caused the original NHS Covid-19 app to be cancelled has been solved. That product often failed to detect cases involving two iPhones because of restrictions imposed on third-party software by Apple.
The team behind the new app acknowledges more work needs to be done to reduce the number of false positives and false negatives that occur at 2m, but is optimistic they can achieve this.
Part of the problem at present is that Apple and Google refuse to share the raw Bluetooth signal data involved.
While the two show no signs of backing down, they will shortly release a new version of their tool that should improve matters.
This development has also been welcomed by those involved with Switzerland’s SwissCovid app.
“While the updated Google/Apple exposure notification API [application programming interface] still aggregates and shuffles data for privacy reasons, it will expose more information needed by the app to compute exposure more precisely,” explained Prof Mathias Payer from the EPFL university in Lausanne.
Test and Trace officials say the motivation for the app is to give “maximum freedom at minimum risk”, but acknowledge it is not a “silver bullet”.
“By launching an app that supports our integrated localised approach to NHS Test and Trace, anyone with a smartphone will be able to find out if they are at risk of having caught the virus, quickly and easily order a test, and access the right guidance and advice,” said Baroness Harding.
However, she is not yet ready to say when a national rollout could occur.
An academic who had served as an ethical advisor to the original scrapped app was positive about the fact that the trial was not limited to the Isle of Wight this time.
“This time it’s a more diverse area – and not just one full of older white people – because it was clear that before very little could be gained from analysis of the demographics” said Prof Lillian Edwards.
But she added that the government still had a “battle to persuade people” to install the software.
“The evidence from Italy is that people aren’t installing their Immuni contact-tracing app, but they might when the number of infections rises again.”
Another public health expert was even more sceptical.
“Even if they have got it working, the app is unlikely to make a difference,” said Prof Allyson Pollock from Newcastle University.
“The issue is not just the contact tracing but the ability to get people to isolate and quarantine. And that means financial support needs to be provided by the government.”
US President Donald Trump issued executive orders on Thursday banning any US transactions with ByteDance, the Chinese company that owns video-sharing app TikTok, and Tencent, owner of the WeChat app, starting in 45 days.
The orders come as the Trump administration said this week it was stepping up efforts to purge “untrusted” Chinese apps from US digital networks and called the Chinese-owned short-video app TikTok and messenger app WeChat “significant threats.”
The TikTok app may be used for disinformation campaigns that benefit the Chinese Communist Party, and the United States “must take aggressive action against the owners of TikTok to protect our national security,” Trump said in one order.
In the other, Trump said WeChat “automatically captures vast swaths of information from its users.
“This data collection threatens to allow the Chinese Communist Party access to Americans’ personal and proprietary information.”
The order would effectively ban WeChat in the United States in 45 days by barring “to the extent permitted under applicable law, any transaction that is related to WeChat by any person, or with respect to any property, subject to the jurisdiction of the United States, with Tencent Holdings Ltd.”
Trump said this week he would support the sale of TikTok’s US operations to Microsoft Corp if the US government got a “substantial portion” of the sales price but warned he will ban the service in the United States on September 15.
