Apps
Tinder looks to revamp as it sees new apps are eating into their market share
…data shows that users are opting to sign on to new apps following a return to the dating scene post-pandemic. It’s still by far the dominant app in the market which is a major advantage for dating apps as users would naturally seek for the pond with the most fish. You would think therefore that Tinder should be in a comfortable position – not so; 75% of users are men with more women opting Bumble in which only women are entitled to make the first move. You can see how Tinder might easily slip down the rankings.
It’s time to innovate, says the company. Ideas include metaverse dating (which is fraught with issues I think as it could become a hotbed of older predators preying on the young, easier to deceive and to defraud), opportunity to buy virtual coins and gamification. The firm has hired personnel from gaming companies.
Artificial Intelligence/ Algorithms
Google Alphabet partners with Everyday Robots to launch robots that can understand better, called, PaLM-SayCan
…this is best understood by clicking the link and watching the 1.5minute video.
Robots can now understand our intentions – if you say I’m hungry – the robot can interpret that to mean can you get me something to eat? What the robot does is to look up the various commands it can do and chooses one (eg. open food drawer and bring a snack) that is likely to satisfy the commander’s needs. The technology is developed using Google’s large language model (LLM) PaLM.
Google’s DeepMind can now create videos from a single frame – The Transframer
…Just as we had discovered and marvelled at Dall-E and Imagen, which is AI that can create images from written descriptions. One Twitter user proposes to combine Dall-E with Transframer.
DeepMind said on Twitter: “Transframer is a general-purpose generative framework that can handle many image and video tasks in a probabilistic setting. New work shows it excels in video prediction and view synthesis, and can generate 30s videos from a single image”.
This has the potential to minimise costs on content making (be it ads, movies, gaming and metaverse/VR). This is because the AI can generate predictive depth and perspectives. It would have the potential to enhance the world of VR.
Going back to text to image AI- TikTok has launched “AI greenscreen”. Now content creators can type in the background of choice and TikTok generates it for you. It is reported that it looks more like abstract painting which may be because the tool could easily become abused – a background of “Donal Trump being assassinated” – for example. Check out the link to see what that looks like should you ever have a bored moment at work.
Meta uses algorithms to fire contract workers that were supplied by Accenture in the US
…so it is reported. The workers are allegedly told that they were selected at random. In Europe (which includes the UK), you have the right not to be subject to a decision based solely on automated means per Art 22 GDPR (and its implementing laws – for example in the case of the UK).
Capitol Records drops world’s first AI generated Rapper FN Meka to sign up to a major label for the N-word outburst
…and apologised to the black community. FN Meka’s voice is human based but the music is AI-generated. The artist boasts more than 500,000 monthly Spotify subscribers and more than 1bn views on its TikTok account where his incredible lifestyle is posted and NFTs sold.
BigTech/ Data / Platforms
Security
Former head of security Peiter Zatko files a whistleblower complaint accusing Twitter of failing to take basic security measures to protect its user data
…The complaint says that Twitter is in breach of 2011 FTC settlement agreement. Allegations include not deleting information when users delete their accounts, using out of date software, withholding of breach incidents, and there is an allegation that some of the staff actually work for foreign intelligence. The prospects of whistleblower complaints (in which the whistleblower could profit personally) means businesses must operate on the basis that any wrongdoing could well go public, even if it may seem unlikely to be discovered. Twitter vehemently denies the allegation.
Could this be a life line for Elon Musk who desperately wants to escape from having to buy Twitter for ~$43billion as promised…?
Privacy
Website gets fined for GDPR breach by using Google Fonts in Germany
…how did the website breach GDPR just by using Google Fonts? According to Wikipedia, Google Fonts “is a computer font and web font service owned by Google. This includes free and open source font families, an interactive web directory for browsing the library, and APIs for using the fonts …”
How did the use of Google Fonts breach GDPR?
This is how it is explained by one website (click the title link):
This occurs even if the User refuses to consent for its IP address to be shared. Because it is possible theoretically to identify a person by its IP address, this qualifies as personal data. As the US does not have GDPR adequate rules on personal data protection, the website thereby breached GDPR by sending the IP address information to the US.
The website has fined €100, but it faces a fine of €250,000 for each violation, or up to six months in prison, for continued breach by the use of Google Fonts. It is though possible for the website to save the fonts locally and integrate them into the website, rather than use the Google Fonts API – which triggers the particular font to be fetched from the remote server as the website loads.
Google’s automated scanning and analytics results in the erroneous reporting of a man to police as child abuser and wipes all data off his Google account
…all the man was doing was helping his child receive medical care concerning the groin area by taking images to show the doctor. The photos were automatically updated to Google’s cloud, which identified the image as a CSAM (Child Sexual Abuse Material). The man has been cleared of all charges but his account (including his emails) which had been wiped off on the basis that the user breached Google’s policy, will not be reinstated, Google says. This is why ultimately, humans need to be involved to decide whether it is a CSAM or not – and Google says its staff are trained by medical experts. But there is a limit to human ability, and humans do make errors. In many ways, such an incident is unsurprising, if one thinks about how many reviews Google needs to conduct.
The UK government, by the way, has clarified how service providers should make judgements about content on their service, including whether or not it amounts to illegal content and must be removed in relation to the upcoming UK online safety bill by adding further amendments. The new clause establishes that providers’ systems and processes should consider all reasonably-available contextual information when making judgements about whether content is of a particular kind. In light of the Google mishap, this clarification is much needed – but the increase in burden on service providers may be significant.
Oracle faces class action for violation of the privacy of “billions of people across the globe” – Northern District of California
Lawsuit complaint says: Oracle, in the course of functioning as a worldwide data broker, has created a network that tracks in real-time and records indefinitely the personal information of hundreds of millions of internet users. Oracle sells this detailed personal information to third parties, even where those persons have no direct relationship with Oracle. It also jeopardises the privacy of those seeking abortions in states where they are illegal.
What does Oracle do?: The suit describes Oracle as the world’s largest data broker (including personal information), a database related software provider and a data storage service provider, including “Oracle Cloud” that developers may use to build and run internet sites and mobile applications. The data brokerage business includes racle Data Marketplace which is one of the world’s largest, commercial data exchange and Oracle ID Graph is a service product designed to provide “identity resolution,” the process of “matching individual customer identities . . . and combining them into a single consistent and accurate customer profile.” This synchronizing allows Oracle to identify individuals and aggregate their many identifiers. This ID enables the marketer to orchestrate a relevant, personalized experience for each individual across marketing channels and device types. Oracle and other data brokers act as central nodes in the “adtech” network…and used to identify and profile individuals for “targeted advertising” or other commercial and political purposes. Oracle is also said to facilitate the sale of sensitive data, such as race, location, politics, medical profiles in its marketplace:
How does Oracle collect data?: According to the suit, Oracle collects many types of personal information from Internet users including concrete identifiers such as names, addresses, e-mail, and telephone numbers. Oracle also amasses data about peoples’ behavior, including the sites they visit online, their digital and offline purchases, where they shop, and how they pay for their purchases. Oracle gathers this personal information from a suite of its own Internet technologies, including cookies, tracking pixels, device identification, cross-device tracking, as well as from its acquisition of data from other parties. There is further detail on how Oracle is alleged to use a piece of software to track user’s behaviour including intercepting communication with websites. The complaint also says that Oracle bought Datalogix, an information broker which purchases and aggregates data from retailers’ loyalty programs.
Cause of action: None of the data subjects have consented to data collection or use of personal information in breach of Californian privacy laws, unfair competition laws, wiretap act, unjust enrichment, and other laws.
Interesting alleged fact: the suit says – Oracle takes its name from a CIA project codename. In 1977, Oracle’s founder, Larry Ellison, was hired by the CIA to build a database; the CIA was Oracle’s first customer. The complaint says surveillance is in Oracle’s DNA…
As an aside: Oracle is now starting to audit TikTok’s content moderation processes to ensure the Chinese government is not manipulating US users.
Business
Amazon to embark on TikTok like feeds
…Everyone is mimicking TikTok, it’s super successful algorithm seems to have successfully eaten into a number of tech giants’ lunch. That is pretty amazing. Everyone else – YouTube, Meta, Twitter etc seem to have tried to copy TikTok in a bid to catch up. The impressive thing about TikTok is that it remains very viable despite that (Note: TikTok in 2021 became the number one internet domain overtaking Google, according to one report). Compare that for example, Zoom x Microsoft Teams – Zoom was really the first to get easy videoconferencing on the market (although that there were other products before Zoom), and then Microsoft Teams come in to provide a wider, integrated service bundling with other products (eg. Office 365) to make the offering more palatable costs wise.
But now, Amazon is reported to be trying its hand on getting into TikTok like service, albeit in a narrow way it seems. It is understood that they are internally testing a social feed (called Inspire) which may be described as inspired by TikTok. Because Amazon has better data than TikTok when it comes to purchasing behaviours, it has the potential to be very effective. On Inspire, users are presented with items which they might want to purchase.
…but even if Amazon were to fail, that’s OK. According to the linked report, the tag #amazonfinds has 25.6 billion views on TikTok. Influencers have earned millions by recommending products, on which they would earn a commission if users click on the link. By Inspire, Amazon aims to bring some of those recommends directly on its platform. With inflation rising and energy costs surging, and consumer belts tightening, Amazon would wish to eke out as much revenue as it possibly can to satisfy shareholders’ expectations.
Amazon pitches to buy tech based home health provider Signify Health
…Signify Health offers a platform which relies on analytics to provide “value-based” payment programs – meaning it will be paid based on health outcomes [a concept well known in the healthcare sector]. Such a model prevents indiscriminate selling, such as selling services which you know will not particularly improve a person’s health. There are other bidders for the company such as CVS and UnitedHealth.
The motivation for Amazon’s bid is obvious. It has recently moved to buy One Medical, a US primary health care provider (provision of doctors and nurses to provide medical services at centres). It can utilise the data obtained from one service to the other to great effect, it will mean deeper datasets for better service [ie: critical if you are operating a business based on value based payment] and even better network effects [more patients, more data] – plus nobody can beat them in logistics. There could also be synergy with PillPack (now Amazon pharmacy) which delivers pills – the home-health service provider coming to your home to sort you out, can bring the prescribed medicaments with them, explain how medicaments need to be taken (eg. in the case of patches, or administering injectable medicaments). It can then be offered on tiered Prime service subscriptions [this is just one idea of my own, as a possibility].
With all that health data, nous in artificial intelligence, super computing power, I would not be surprised if Amazon were to one day embark on drug discovery…
But on the other hand it has announced that it will shut down Amazon Care – the virtual and in-home health service it initially created for its employees, explaining that it was “not a complete enough offering for the large enterprise customers we have been targeting, and wasn’t going to work long-term”. I find it difficult without more information to make sense of the decision but no doubt there is a good reason for this.
Peloton confirms that Amazon will sell its bikes
…no longer selling through its own channels, it has decided to reach out to millions through Amazon’s network of customers in the US. “We want to meet consumers where they are, and they are shopping on Amazon”, Peloton said, wholeheartedly endorsing Amazon’s reach.
Amazon capitalises on synergy by launching Ring Nation – TV show that features videos taken by Amazon’s Ring device that have gone viral
…it will show neighbours saving neighbours, marriage proposals and funny animal antics taken by the Ring camera. Amazon also owns a streaming service (Amazon Prime Video and a free version called Amazon Freevee) and MGM, the Hollywood film studio. Given that there are no scripts, no sets etc, and limitless scenarios to choose from when one thinks about the gazillion cameras out there that are on for 24/7, budget is cheap, and it has mass appeal.
Amazon to invest in Japanese cosmetics seller Istyle – what might be Amazon’s real objective?
…first blush, [pun intended] there’s nothing extraordinary; unless you know the Istyle’s edge – the enormous amount of comments posted about various cosmetics that they sell. In other words, it has formidable strength in “word of mouth” advertising, and deep understanding of what women really want. These comments contain a lot of intel about consumer desires, sentiments and issues. It would be interesting to understand whether Amazon will supply it with AI know-how to enable Istyle to pick up better signals to improve the business. Note that cosmetics is a high margin, very personal, sticky business, so do it well, and it could translate into more bucks.
Ebay delves deeper into selling collectibles
…they’ve acquired NFT collectibles marketplace KnownOrigin some two months ago, and now they are buying TGCplayer, marketplace for trading cards (eg. Pokemon, football and basketball cards) which can go up in value quickly depending on how rare. A bit like Amazon’s tactics – Ebay are acquiring businesses that can exploit their expertise –ie: marketplace. Just a bit more niche.
Google to launch “helpful content update” to tackle clickbait
…This ensures users sees more original, higher quality content. Content creators can be successful because they know how to optimise the search engines well. That doesn’t mean it’s the best. Google proposes to change that, so that the algorithm picks up more high-quality content. Google says that testing shows that the update has resulted in particular improvements for searches related to online education, arts and entertainment, shopping and technology.
Brain Computer Interface
A paralysed man breaks the record for wearing a Brain Computer Interface – 7 years and 3 months and still counting
…the device enables paralysed people to control prosthetic limbs / other devices such as a computer. FDA has approved Synchron to start clinical trials for Stentrode, which translates brain activity into standardised digital language, enabling paralysed people to text, email etc, by their thoughts. The implantation process is said to be minimally invasive as it does not require the drilling into the skull, unlike other offerings such as Elon Musk’s Neuralink – “a Fitbit in your skull”, which promises to enable the same sort of functioning. Neuralink’s implants will be tested in patients with severe spinal cord injuries this year, following successful demos on animals.
…and now Musk has reportedly approached Synchron to do some kind of deal with Neuralink…
UK Law Society publishes a report on Neurotechnology, law and the legal profession
…it covers off legal implications of neurotechnology developments, such as impact on human rights, mental privacy and surveillance, equity and discrimination, workplace brain monitoring, charging for legal services based on billable units of attention instead of hours.
Crypto
Tokens to help save the Amazon Forest
… Deforestation in Brazil reached a record high in the first seven months of the year. In comes a great idea – if it works. Nemus, a business that offers “a collectible NFT experience” is issuing tokens in relation to parcels of land in the Amazon – to thwart potential buyers of land with intention to knock it all down. The token enables the owner to decide the way in which that piece of land should be protected, with the potential to be awarded with carbon credits. A good idea but Nemus’ plans are getting unstuck because (i) it is experiencing difficulties producing deeds to the land, (ii) the land is very difficult to reach and (iii) despite the promise, a portion of its 41,000 hectares of land already appears to have been deforested.
Other ReFi [Regenerative Finance] projects underway, include the opportunity to protect parts of Democratic Republic of Congo’s land from oil and gas exploration.
EV/AVs
German Court decides Tesla’s “Autopilot” label and “Full Self-Driving” description not misleading
…rejecting a complaint by self-regulating industry body, the ” Center for Protection against Unfair Competition”, which considered that the brand “Autopilot” label for just an advanced driving assistance system (ADAS) gave the wrong impression that vehicles are fully self-driving – which is not legal in Germany at present. Earlier this month, the Californian Department of Motor Vehicles (DMV) made a similar complaint against Tesla.
Tesla to raise the price of its assisted drive system (the Full Self Driving as referred to above) by 25%
…from $12k to $15k for the US which will be offered the updates first. The new update provides driving smoothness, better lane centering and changing, better turns, tackles false slowdowns, creeping for visibility at intersections, better speed when entering highways. Musk says it’s a major code change. Drivers wonder, is it worth it? It is unknown whether its monthly Full Self Driving (FSD) service subscription will increase from the current US price at $199.
For me, the interesting point is that you buy the hardware, and you are expected to continually pay for software updates. The current update on offer does not appear to impinge on security. But if it were to, more people will be more compelled to update. Will this be the way forward for other types of hardware such as smartphones?
Green Technology
Heliogen and Dimensional Energy to produce carbon-free Sustainable Aviation Fuel (SAF) from Concentrated Sunlight and Green Hydrogen
…the “Letter of Intent” (meaning it’s just an intention at this stage) says it aims to create jet fuel from solar energy and air. Heliogen says that they “will work to deploy Heliogen’s proprietary, artificial intelligence (AI)-powered HelioHeat™ technology to convert sunlight directly into thermal energy in the form of high temperature steam and air that will be used to produce green hydrogen for Dimensional Energy’s Reactor platform”.
Bill Gates-backed Heliogen concentrates solar rays by using AI. HelioHeat can also be used for power heavy industrial processes including the making of cement, steel, and petrochemicals.
Other SAF news includes that of Synhelion. The website announces that airlines SWISS and the Lufthansa Group have agreed to use its solar aviation fuel. This will make SWISS the first airline in the world to use ‘sun-to-liquid’ fuel. The process devised by Synhelion uses concentrated sunlight to produce carbon-neutral kerosene. Note that, compared to the Heliogen/Dimensional SAF, this one is not carbon-free but uses less carbon than traditional jet fuels.
Metaverse
Mark Zuckerberg’s selfie with the Eiffel tower and Sagrada Familia to celebrate the launch of Horizon Worlds in France and Spain gets mocked for looking too basic
…Zuckerberg responded with a far more graphic version, explaining that it was taken “very quickly” to celebrate the launch. Scepticism for the viability of Metaverse is high yet Meta has invested billions into the potential. That is why (and also the reduction in ad dollars – partly as a result of the economic slow down and partly because Apple has withdrawn Meta’s access to user data to enable them to target-ad more accurately) the latest launch was clearly a cock up. But it’s nice to know that sophisticated businesses like Meta can make basic mistakes too. Bring out the goods, and it will be forgiven for it.
Supply Chain
Apple Watch and MacBook laptops to be made in Vietnam
…the fact that Apple Watch – a sophisticated product – is going to be made in Vietnam demonstrates a stamp of approval from Apple of its capability. The move from China to elsewhere is much expected, when one thinks about the geopolitical tensions but even aside from that in my view, there may have been a rise in labour cost as Chinese capability improves, which may have played a part in moving the needle.
…Apple is also bringing forward plans to set up a manufacturing site in India – for iPhone 14. These moves will certainly not go unnoticed by the Chinese government.
Delving Deeper
UK issues regulatory plan for self-driving vehicles for rollout by 2025 and publishes a consultation on safety ambition
…the UK is perhaps more cautious ut not necessarily behind the superpowers US and China which already have fully self-driving cars on the roads. The publication sets out how it will realise its vision for Connected and Automated Mobility (CAM) for the UK. At the same time a focused consultation on how adequate safety ought to be achieved, was published. The government’s focus is on the three following “pillars”
- Safety and Security
- Legislating for safe self-driving making clear who/what is responsible for what- It is the government’s intention that this primary legislation will create new legal actors and provide powers for new processes, including authorisation and in-use regulation (see pg. 45 of the Plan).
- Ensuring vehicles are safe, including from a cyber security perspective
- Facilitating safe trials – UK offers 6 CAM Testbeds, which is an environment for the modelling, simulation, testing, and trial deployment of connected and automated mobility solutions.
- Address public concerns, win confidence
- Securing the industrial and economic benefits of CAM
- Delivering jobs and investment – Identify areas of expected UK competitive advantage in the future global CAM supply chain.
- Government funding
- Encourage inbound investment
- Build skills and understanding of CAM technologies at all levels, from local to central
- Co-ordination across government and industry bodies
- Delivering the Societal benefits of CAM
- Decarbonisation
- Integration of CAM into wider networks such as road networks, mass transit systems and wireless network – ambition to ensure the “majority” of the population has access to 5G by 2027.
- Reduce road congestion, increase efficiencies for freight, improve public transport
Some definitions
- Self-driving vehicle is one that has at least one self-driving feature, delivering sufficiently high levels of automation that meets a legally defined threshold and is capable of safely driving itself with no human input. Such features could provide self-driving capability for all or part of a journey. It is intended that ‘self-driving’ becomes a protected term for the purposes of marketing products to the public. [Tesla will need to change its branding in due course, then – see the above news on the point]. There will be two types:
- Vehicle that can drive for the entire journey, humans would be mere passengers. Referred to as No User-in-Charge (NUiC) vehicle. Requires a licensed ‘NUiC operator’ (which could be the same as ASDE – as to which see below) which would be responsible for overseeing the non-dynamic driving task responsibilities, such as ensuring the vehicle has appropriate insurance, that would otherwise remain with the User-in- Charge
- Vehicle which can drive under certain conditions (eg. day time, major roads only). It will require a Human User-in-Charge (UiC) for driving outside those conditions.
- Automated can refer to a wider range of automation, including technologies which are not capable of self-driving. While the term ‘automated’ vehicle will continue to be used by the sector and in legislation, ‘self-driving’ and ‘self-driving feature’ are better terms to support public understanding.
Other noteworthy points
The plan is quite long. Here are some key concepts worth noting though (page 43 of the Plan – see the link – is quite useful):
- Automated and Electric Vehicles Act (AEVA) 2018: Trialling of a self-driving vehicle with a safety driver is already possible on any UK road. Once a vehicle no longer requires a safety driver it may meet the definition of a self-driving vehicle, as set out in this Act.
- How safe? Government proposes that it ought to be the same standard of behaviour as that expected of human drivers; competent and careful. This strikes the right balance, the government says and it’s a higher standard than the average driver which includes fatigued, distracted or under the influence of drink or drugs [!! – is that really the average driver in the UK…one hopes that the latter two are outliers…] – too high and it will stifle innovation too low and the benefits are eroded. [Contrast with the quote of Elon Musk: “Being better than a human is relatively straightforward, frankly…but how do you be 1000 percent better, 10,000 percent better? That’s much harder” – not a criticism of the government for proposing to set the standards lower though, but it’s a critical question, to which the government invites views in its consultation]
- Authorised Self-Driving Entities (ASDEs): They will be responsible for the behaviour of self-driving vehicles. They should be under the duty to report any discrimination – vehicles that discriminate against minorities or vulnerable people. The level of discrimination is then assessed for acceptability. An ASDE will need to be registered. An ASDE is likely to be a vehicle manufacturer or software developer, or a partnership between the two.
- Safe by design: vehicles to be assessed for safety at different points of the vehicle’s lifecycle. Level of detail required to satisfy the assessment will be determined following the consultation.
- Automated vehicle approval: technical assessments of each self-driving feature to be carried out as part of the automated vehicle approval.
- Authorisation to self-drive: Assessment to check that vehicle can self-drive (without human monitoring) in various circumstances. Authorisation will also consider whether a suitable ASDE can vouch for the safety and lawfulness of the vehicle. The vehicles will be subject to a monitoring test which have 5 criteria:
- Compliance with relevant road traffic rules
- Avoidance of collisions which a competent and careful driver could avoid
- Avoidance of causing collisions
- Treatment of other road users with reasonable consideration
- Recognise when it is operating outside of its operational design domain
- In-use regulation: This ensures that the vehicles continues to fulfil the safety criteria once on the road. Data would need to be fed to the “in-use regulatory scheme” by ASDEs and operators of the self-driving vehicles (NUiC operators). [Data would need to be in sufficient detail to explain what happened in the case of an incident]
- Cybersecurity consideration: International cyber security regulations for vehicles (set at the UNECE, or United Nations Economic Commission for Europe) to apply. Government to consider whether additional requirements should apply. See also National Cyber Strategy 2022 which sets out the aim to build on the National Cyber Security Centre (NCSC)’s security principles for connected places to reduce the risks.
One More Thing…
Elon Musk lookalike Yilong Ma is back – on Instagram!
…after being booted out of Douyin (ByteDance’s TikTok offering in China) and Weibo (sort of China’s answer to Twitter) in May, he is back and attracting attention. Elon himself has set he’d like to meet him, but at the same wondering whether he could well be a deepfake. Click on the link to see what deepfakes could be like these days. It’s very real – and has much comedy value.