Archive

Headlines in Tech News of the week

Nvidia and AMD ordered to stop exporting cutting edge machine learning chips to China

…the news broke as I was issuing last week’s digest, so it’s a week old now, but there has been some further developments and rumours…

US government has imposed a new license requirement, effective immediately, for any future export to China (including Hong Kong) and Russia of the companies’ cutting edge AI chips to China. In particular, what is being banned is electronic design automation software which are needed to design high-end chips. System designed to integrate those are also affected. Chinese companies affected are said to be mainly cloud service providers, internet and AI companies.

The US government says the new license requirement will need to address the risk that the covered products may be used in, or diverted to, a ‘military end use’ or ‘military end user’ in China and Russia. This means Nvidia and AMD stand lose if they cannot fulfil their contract with their Chinese customers. The government has since given Nvidia the green light on developing one of its AI chips in China, and to continue exporting AI chips from its Hong Kong facility to 2023 [but this would seem to defeat the purpose of the export ban].

US is rumoured to step up the restrictions on Chinese companies by limiting how Chinese internet companies can collect data on U.S. citizens and set out the technologies that cannot be exported to China.

Learning point from this is that, businesses will be in the future required to think about possible regulatory pressures in advance. It seems a bit unfair for the regulation to come down on the chipmakers so suddenly. It may be wise if contracts could address what happens if there are sudden tightening of regulations which means that the promised supply cannot be fulfilled through no fault of the supplier – this will all depend on the relative bargaining position of the parties. 

It’s a tricky one because we just don’t know where the line lies. If not cutting edge machine learning chips, what about not-so cutting edge machine learning chips?  What about other products that could possibly be used for military use, but has plenty of civil uses?

Artificial Intelligence/ Algorithms

Promptbase – a marketplace for AI art created by text to image AI engines such as Dall-E

…now everyone can make and sell art.  But what stops people from trying to get something similar using Dall-E or other, text to image AI? There is still skill required in editing the AI generated art though, and complicated art will be difficult to replicate even where it is generated by AI, at least for the moment.  

Interesting copyright questions are likely to arise. What is the copyright status of all the data which was scraped to make the AI engine the first place? What does that mean for the generated AI art?

Dall-E says users get full usage rights to commercialize the images they create with DALL·E, including the right to reprint, sell, and merchandise – but Open AI claims to own the original image. Other text to image generating AI have their own terms; including a stake in the proceeds should the art be onwardly traded.

Where the terms as to ownership are unclear, the outcome may depend on the law of copyright, which may differ depending on the jurisdiction. In the UK, it is arguable that AI generated by copyright is protectable as a computer-generated work under s.9(3) Copyright, Designs and Patents Act 1988 (CDPA), in which case the author/owner will be “the person by whom the arrangements necessary for the creation of the work are undertaken” [there may be argument over who actually undertook the arrangements necessary]. The EU position may be different, as case law indicates that human creation is required to attract copyright (eg. Case C-145/10 (Painer), Case C-683/17 (Cofemel)). In the US, the argument is likely to centre around the scope of fair use – which is a legal doctrine that promotes freedom of expression by permitting the unlicensed use of copyright-protected works in certain circumstances.  

Worth checking out is Dall-E’s new feature, Outpainting “which helps users extend their creativity by continuing an image beyond its original borders”. Check out the Girl with a Pearl Earring in her surroundings, here.

US Department of Defence opens a marketplace Tradewind – to find, fund, and accelerate the adoption and transition of AI technologies to solve DoD challenges and problems

… The marketplace is led by the Pentagon’s Chief Digital and Artificial Intelligence Office (CDAO)

Meta uses AI to decode speech from brain activity

…no need for transplants, and gives those with brain injury/ stroke victims so severe that they cannot type to communicate, much hope. Facebook AI Research Labs (FAIR) scientists have been able to show “from three seconds of brain activity… [the ability to] decode the corresponding speech segments with up to 73 percent top-10 accuracy from a vocabulary of 793 words, i.e., a large portion of the words we typically use on a day-to-day basis.”

Reddit undertakes AI acquisitions for Ad-targeting purposes in amid data use restrictions

…Last year Apple gave iPhone users the option to stop third party apps tracking what users do with their mobiles phones (such as Apps and Websites). Tracking enabled those third parties (for example, supermarkets, car makers, fashion brands – you name it) gather data about users and understand their likes and dislikes. That was then used to target ads – car makers could spend cash on advertising on eg. Facebook only on users that liked cars, Gucci might advertise on users that had more expensive buying habits, or bought other luxury brand items, for example. The users took up the option to block this commercial surveillance with zeal such that brands could no longer ad target effectively, and advertising space providers such as Facebook and Reddit experienced a cratering in its revenue. Added to that are ever tighter regulatory restrictions on the horizon to ensure users’ data are not taken and exploited without regard for their rights, privacy or security.

Reddit is now on a shopping spree to buy AI technology that enable it to glean the conversations on its platform to better understand trends – so they can better ad-target, and so reap higher income from advertisers now that data is less available. It can also use the technology to content moderate, provide a better service to users by enabling them to get to the most relevant content (without having to develop super indexing algorithms like Google), and perform on point automated recommendations (like TikTok). It has so far purchased:

• Spiketrap – a company that uses artificial intelligence to contextualize audience understanding
• MeaningCloud – a natural language processing platform that enables developers to build apps that can extract meaning from written content.
• Spell – platform for running machine learning experiments

Who is liable? Truck drivers subcontracting for Amazon blame Amazon’s AI for poor working conditions in Japan

…the unionised truckers say the AI diverts them to overly narrow roads [lots of narrow roads in Japan], and doesn’t take into account driving conditions [such as weather]. This increases their working load. The truckers are subcontractors and technically work for a third party logistics company so Amazon says it is not legally responsible. The union for the truckers say they get the orders from Amazon through the Amazon App, so they work for Amazon, who should be responsible. Employment lawyers, take note – this issue may arise in other jurisdictions.

BigTech/ Data / Platforms

Content moderation

Cloudflare, the internet infrastructure company decides to remove support for Kiwi Farms – a notorious hate speech site

…Cloudflare protects websites from DDOS attacks (bombarding of traffic on target websites – think website crashes when in demand tickets are released). The threats on the site escalated enough in the last 48 hours that, in spite of proactively working with law enforcement, it became enough of an imminent emergency we could no longer wait for them to act – said Cloudflare CEO. The site was particularly known for harassing and doxxing (publishing malicious private info) trans people.

Competition

Apple caves and settles with App Developer Flicktype

…the complaint concerned an activity known as Sherlocking. This refers to a time when Apple sunk a popular App priced at $29 called Watson by later releasing its own finder app with capability of carrying out a search on the web and files on the device – called [aptly named?] Sherlock. Since then Sherlocking came to describe the move where a popular app becomes redundant because it becomes integrated within the operating system. The idea is not dissimilar to Amazon using non-public data of sellers on its platform to inform itself of which best-selling products it should focus on selling on the same platform – the EU Commission is currently investigating whether such practice is anticompetitive. 

Flicktype was an app offering keyboard for the Apple watch. The suit says that Apple was interested in buying the company, but then the App was told that it needed to be removed because the app is a keyboard for the Apple Watch. Apple then launched its own swipe keyboard app alongside the new Apple Watch Series 7. Reportedly, there are a number of instances of Sherlocking by Apple. As pointed out by FOSS patents, the better course may be to see if the functionality of the apps can be patented.

The settlement with Flicktype is significant because there are rumours that the US Department of Justice is preparing an antitrust suit against Apple for “abused its market power to stifle smaller tech companies, including app developers and competing hardware makers.”

This again is perhaps an example of BigTechs making moves before antitrust suit is filed against them. Recently, Apple has enabled users to delete wallets from iPhones, and will allow third party web browsers to work with Apple Pay – following a complaint made to the EU Commission. Last year, Apple enabled a Report button enabling users to report scammy apps – which may have been due to the up-coming EU’s Digital Services Act which oblige “very large platforms” to analyse systemic risks they create and to carry out risk reduction analysis to reduce risks associated with, among other things, dissemination of illegal content.

Platforms

Apple not liable for making available phishing app Toast Plus on app store says Judge (Northern District of California)

…interestingly one of the class action plaintiffs was not a resident in the US, but Japan (who lost out big time by the scam). Plaintiffs downloaded a scam e-wallet app which stole cryptocurrencies. Plaintiffs believed that Toast Plus was a version of Toast Wallet, a well known cryptocurrency wallet, as the names were similar and the logo used for the application in the App Store was the same or nearly identical.

The plaintiffs point out that Apple is representing in its campaign (eg. against antitrust complaints concerning App Store monopoly – ie: not allowing third party apps on the iOS ecosystem) that App Store is “a safe and trusted place” and that “that the apps we offer are held to the highest standards for privacy, security, and content”, yet it had listed Toast Plus, a fraudulent app, on its app store.

That was not enough to make Apple liable, the Court said. It was immune from liability arising from content created by third parties under § 230(c)(1) of the Communications Decency Act (“CDA”).  It was a publisher in that it reproduced a work intended for public consumption, the third party content for which Apple did not have any hand in developing. Claims under consumer protection laws failed, and Apple was able to rely on its terms limiting liability.

Amazon to deliver prescription pills in Japan by partnering up with pharmacies

…Amazon has recently invested (not acquired) in iStyle, a Japanese cosmetics company, with a rich source of online discussions. In Japan, a relatively wide variety of cosmetics are sold in pharmacies. Perhaps there is some sort of synergy with that business too. Japan is a lucrative country for pharmaceutical companies as a significant portion of the population prefer originator drugs to generics. That’s another reason for expanding into Japan. Further, it may be trying simply to grow internationally. The antitrust climate in the US is pretty hot, so in order to increase revenue and perhaps access to data – which can aid business to be sharper if data is processed right – it may be expanding into other countries.

Separately, on a health related business point, Amazon was outbid to buy Signify Health, value-based care and provider enablement business in the US. CVS will be buying it for $8billion. Amazon may have lost the bid, but it may have done some collateral damage to would be rival CVS (if Amazon gets the go ahead to buy One Medical) by pushing up the acquisition price tag.

Privacy

Instagram fined €405million by Irish data regulator for not keeping children’s data safe in breach of GDPR

…Instagram explains that the breach concerns an old feature in which all users, including under 18’s default setting was not private when it is changed to a business account, as a result of which childrens’ data (eg. phone numbers, emails) were automatically disclosed. That has now been remedied. In any event, Instagram said they will appeal. #

UK has been leading the efforts to protect childrens’ data with the Children’s Code, with many countries following that lead with stringent rules obliging businesses to keep childrens’ data private. California has last week approved age-appropriate design code for certain sized businesses, which includes privacy by default settings.

Crypto/NFT

Proceeds of Actor Bill Murray’s charity NFT is stolen – Ethereum worth $185,000 …the interesting point is the use of the NFT with a real world element. Bill Murray auctioned “Beer with Bill Murray” NFT – the winner got to have a beer with the actor, the scene of which an artist painted on canvas and then converted into a digital copy.

Cyber Security

Criticism over decision of Lloyd’s of London to exclude state-sponsored cyber attacks in insurance policies

…Lloyd’s say, it is (i) akin to exclusions for acts of war and (ii) owing to the scale of potential losses it could make policies prohibitively expensive.

Critics say state backed attacks are so frequent that businesses might decide not to buy a policy. Policy prices have increased sharply in the recent year.

EVs/AVs

UK Data Protection Regulator prosecutes theft of data from Vehicle Repair Garages

…voluminous data of 100s of thousands of people relating to those involved in car accidents were stolen to obtain leads to make insurance claims. They were prosecuted under the Computer Misuse Act 1990, which carries a custodial sentence of up to two years.

Uber partners with fintech Moove in London

…London has ever burgeoning strict and expensive low emission zones. These leave EVs out of reach for most drivers who would work for Uber, who has the ambition to become a 100% electric platform in London by 2025. To drive under the UberX banner, cars must be at least a hybrid.

Nigerian fintech Moove, which has an established presence in the African and Indian regions, proposes to help with the financing to add 10,000 extra Uber cars on London’s roads to rent out to drivers on a weekly basis. Drivers need not undergo credit checks or a deposit, instead preferring to use alternative credit scoring technology which is expected to be more financially inclusive. Moove will also provide Moove Charge, an app for Uber drivers, giving them access to 6600 charge points.

A day after GM’s Cruise gets Robotaxi (no driver) licence in California, it causes an accident – leads to software recall and update in 80 Robotaxis

…not the best timing, but the accident was not fatal. The Robotaxi failed to avoid an oncoming traffic speeding in the wrong lane – 40 miles per hour in a 25-mph restricted right-turn/bus-only lane. As the Robotaxi started to make a left turn, it predicted that the oncoming traffic would turn right – and so it hard-braked. Instead, the car drove straight on and the two vehicles collided.

Query I have is what percentage of human “competent and careful” drivers [to borrow the standard the UK Government is proposing to use] would have been able to avoid it.

Metaverse / VR / AR / MR / XR

Meta enters into a “multi-year broad strategic agreement” with Qualcomm to build custom chips for Meta’s VR headsets

…think I heard that one before…. Qualcomm recently entered into a multiyear broad strategic partnership with Samsung. And what do Samsung and Meta have in common? A common competitor in the name of Apple – in the mobile (Samsung) and VR (Meta) space.  Apple itself has locked horns with Qualcomm, unhappy with Qualcomm’s pricing and pricing structure for its chips and intellectual property, leading to a 2019 settlement following a global battle that ended with Intel being pushed out of the chip-making market [though Intel is now entering back into it]. The leftover skirmishes of that battle ended about two months ago. Apple itself is doing all it can to reduce its dependency on Qualcomm, by designing own chips [it’s forthcoming anticipated VR headset to be based on in-house designed chip, the M2] and getting the top manufacturer TSMC to make them. Qualcomm/Apple licence ends in 2025.

Product and engineering teams will deepen technical collaboration to develop next generation Snapdragon XR chipsets. Meta’s current VR headset, the Quest 2 uses Snapdragon XR2. Notable points:

• Chip will be optimised for Meta’s system specifications, but will not be exclusive to Meta.
  • Note Qualcomm has a partnership with ByteDance [TikTok parent] who owns VR headset company Pico, and also working with Microsoft on custom chips for next-gen AR glasses.

• Meta is trying to custom silicon of its own for its VR/AR headset. Mark Zuckerberg is reported to have said “Unlike mobile phones, building VR brings novel multidimensional challenges in spatial computing, cost and form factor”

• The agreement covers only VR devices.

Note

AR (where information is overlaid over the physical world) is much more difficult than VR in terms of technology, which is why VR is being launched first. Companies will then probably move onto MR (mixed reality) – where there is an out-ward facing video camera so the real world can be blended with the virtual (apparently called “pass through” – so it’s not really AR) – you do not see the real world direct, but the processed videoed images.

Lenovo to sell platform neutral AR glasses (which looks like ordinary sunglasses)

… Lenovo Glasses T1 will be usable with Windows, macOS, Android, or iOS devices.

Semiconductors

India moves in on the chipmaking act!

…We’ve seen the US Chips and Science Act and the EU Chips Act with $52billion and €40billion worth of subsidies respectively. Now India is offering $10billion worth of incentive to lure chip makers to manufacture there too.

As the FT points out, as a result of US blockade of exportation of cutting edge chips into China, China is also likely to provide some sort of semiconductor package.

And so is Micron in the US

…it will be the first memory semiconductor manufacturing plant by a US company over the last 20 years. The company says that the support provided by the US Chips and Science Act enabled to embark on the venture. The chips will be made for multiple markets including data centres and autos, where Micron chips are currently being widely used.

Supply Chain

Amazon to provide low-cost long-term storage that gives sellers the option to store their inventory in Amazon distribution centers

…when it comes to logistics, Amazon is king. Here, what Amazon is really offering, is supply chain resilience.

When supply chain disruptions erupted during the pandemic, Amazon was the business that was capable of routing its cargo, using smaller vessels that can pass narrower channels and better able to utilise smaller ports, avoiding the larger ports that were clogging up. It has own cargo ships and containers, has control over every node across their supply chain, have the predictive logistics capability to manage and prioritise goods, and make distribution as seamless as possible, as explained by CNBC.

The new service, “Amazon Warehousing & Distribution (AWD)” gives third parties the opportunity to rely on Amazon’s logistical prowess to buy the supply chain resilience businesses desperately need. Amazon will also have the opportunity to collect more data in the process, although the freedom to use such data as it sees fit, might be restricted by upcoming regulatory measures. Owing to supply chain issues, some businesses are expected to order more than they need so that they can fulfil their contracts come what may. Some companies – like Peloton have found it extremely difficult to predict future demands leading to excess inventory (so much so that it had to do a deal with Amazon to sell them on its platform). Whatever the case, any storage for stock needs to be available and accessible and devoid of further supply chain issues, such as lack of truck drivers. In Amazon’s world, where there’s pain there is gain – a true mark of success.

Telecoms

Senators urge the Federal Communications Commission to “proceed with an evidence-based rule making” when deciding on whether to allow 5G wireless service to the 12 GHz (1 2.2-12.7) Band

…the concern is interference with incumbent satellite services that serve rural areas. The digital divide can only be closed by using the spectrum efficiently and wisely, senators said.

T-Mobile wins 90% of the licences in the 2.5GHz Spectrum from the Federal Communication Commission auction

…the interesting features are:

• most of the available spectrum in the 2.5 GHz band are located in rural areas
• Mid-band spectrum such as 2.5 GHz can boost capacity and speeds over relatively long distances, heavily used in 5G – T-Mobile is actively utilising the 2.5GHz band (cf 3.7 GHz to 3.98 GHz range, which is also in the mid-band range, is used by Verizon and AT&T which had issues on deployment owing to potential interference with airplane navigation systems)

With remote working, there may be scope for development in rural areas, and less competition. T-Mobile’s plan may be to get into rural areas quick and integrate into the rural network, and essentially be the best at providing good network across the country.

One more thing…

Surgeon successfully completes vasectomy despite power outage with the help of Rivian

…the patient couldn’t reschedule, so he consented to have the surgery powered by the EV. Thankfully the outcome was positive. On a bit more serious note, EVs are essentially stored energy on wheels – it can potentially do a lot more things eg. rescuing people in remote areas.

Headlines in tech news of the week

China agrees to let US audit US listed Chinese companies

…the US had threatened to kick out Chinese companies from its stock exchanges otherwise. Any loss of access to US stock exchange would obviously massively reduce access to the largest public capital market for Chinese businesses. The principle from the US side is that all listed companies should be on an equal footing, and investors should be given the assurance that the numbers disclosed by all public companies can be relied upon and not cooked up.

” Make no mistake though: The proof will be in the pudding,” said Security and Exchange Commission head Gary Gensler. “This agreement will be meaningful only if the PCAOB [Public Company Accounting Oversight Board] actually can inspect and investigate completely audit firms in China”.

Apps

Apple enables users to delete the Wallet app

… back in May this year, the EU Commission said that Apple may be breaching competition laws by preventing third parties from providing a mobile wallet system. A lawsuit was filed in July on behalf of banks complaining that Apple is anti-competitively forcing banks to adopt Apple Pay. iPhones have NFC (Near Field Communication) chips which enables users to pay by tapping on the merchant’s device. Whilst Google enables third parties to access NFC on Android devices, Apple has this chip locked in with Apple’s wallet app. The reports say that in order to tackle the legal issue that has been raised by the EU, Apple will enable users to delete the app, and allow third party web browsers to work with Apple Pay.

Query how many people will actually delete the wallet, given that it does give users a very useful seamless method to make payments for goods.

Artificial Intelligence/ Algorithms

Filtering issues emerge with new uses of AI

A variety of different AI tools are emerging but the issues seem to be similar:

• A company called Sanas attracted criticism for supplying AI software that filters out accents used in call centres to make call centre agents sound white American. Sanas (founded by first generation immigrants) say that callers who call in because they are frustrated with a service, get even more frustrated when they sense that someone abroad is fielding the calls – and the agents get the brunt of the call. Sanas also say that callers can understand the agents better, enabling the service provider to improve their service.  Critics say it ignores and disrespects different cultures that exist in the world.  

• Stable Diffusion – which provides AI image generator that produces text to images like Dall-E and Imagen, has caused deepfake problems as unsavoury, and realistic images of public figures are being shared on 4chan, an anonymous English-language imageboard website (hacking group Anonymous formed from 4chan users). The difference between Stable Diffusion and Dall-E/Imagen is the lack of safety measures to stop the ability to generate deepfakes of public figures. The company is now working with “leading ethicists and technologists”, it said. The overwhelming majority of deepfakes are understood to be pornographic. Implementing measures to stop the generation of pornographic images would be a very good start.

AI generated art work wins first prize in the digital category at Colorado State Fair

…this was not against the rules, but most of the art submitted was not AI generated. It has attracted criticism. But note, the “artist” did work on the art, touching up what the AI spat out.

AI powered camera to detect falls in an old persons’ home detects too many false positives

…but I’m sure it would get better. The AI camera was designed to detect up falls and screams. However, the driver for the camera being installed in the first place was because of reported elderly abuse and neglect. This meant that staff were on constant watch and had to respond to every alarm bell. It also meant that it was time-consuming and reduced the time available for care.

French tax authority uncover undeclared private swimming pools using Google/Capgemini developed AI leading to a windfall in taxes

…Small amount of tax €200 /year need to be paid in respect of property that houses a swimming pool. Using aerial photographs and the helping hand of AI, the tax authority has been able to unearth more than 20,000 pools that have not been declared. Pools must be declared because they contribute to the value of property. Eventually, it may be used to detect whether people are observing a ban on swimming pools should there be one in the future, if France should suffer severe drought as it has done this year.

The tax authority have said that it may then extend to undeclared home extensions, patios and gazebos in respect of which tax also needs to be paid.

BigTech/ Data / Platforms

Content moderation

Trump’s answer to Twitter “Truth Social” fails to satisfy Google Play’s standards

…as a result of which Truth Social, which purports to provide a “vibrant, family-friendly environment” is not available on Android phones.  Google cited content moderation concerns.  It means that Trump cannot reach to 44% of the Android using public – at least via their mobile phone. To what extent would this affect any future elections, I wonder. Trump is as I understand it, still banned from Twitter (hence he has decided to found something like it himself). 

Truth Social is available on iPhones.

Privacy

Wikimedia foundation the nonprofit that operates Wikipedia –  petitions the US Supreme Court to subject the US government’s mass surveillance program to constitutional review

…who knew that the US National Security Agency (NSA) has a program (also known as “Upstream” surveillance) to intercept Americans’ private emails, internet messages, and web communications with people overseas.

Upstream surveillance is conducted under Section 702 of the Foreign Intelligence Surveillance Act (FISA), which permits the government to intercept Americans’ international communications without a warrant – so long as it is targeting individuals located outside the U.S. for foreign intelligence purposes. Section 702 will expire in 2023 unless it is reauthorized by Congress.

In September 2021, the Fourth Circuit held that even though the Wikimedia Foundation provided public evidence that its communications with Wikipedia users are subject to Upstream surveillance, “state secrets privilege” meant that the case should be dismissed. Said privilege allows the government to withhold information in legal proceedings if disclosure of that information would threaten national security. Wikimedia seeks the Supreme Court to review the decision.

FTC sues data broker Kochava for selling US citizen’s sensitive information such as visits to abortion clinics

…the adtech group – who has Uber, McDonald’s and Disney as their clients, is alleged to have sold off citizens’ location data, including abortion clinics, medical institutions, homeless shelters, domestic violence shelters, rehabilitation clinics, and religious centres. The agency has sued because Kochava is exposing individuals to threats of stigma, stalking, discrimination, job loss, and even physical violence. Kochava explains that the data collection has been consented to by the individuals. This action follows other similar types of lawsuits advanced by the FTC, as highlighted previously.

Kochava is in the process of enabling blocking geolocation to mitigate the claim raised.

Restaurants using voice recognition technology sued in a proposed class action for breaching Illinois Biometric Information Privacy Act (BIPA)

…the defendant restaurants have collected customers’ voiceprints, when using the automated voice ordering system, without obtaining the requisite consent / providing customers with data retention and destruction policies per BIPA, says the complaint. If biometrically unique information fall into the wrong hands, it can lead to identity theft, as it can be used to extract sensitive data.  

The complaint goes on to explain: “To the human ear, your voice can instantly give away your mood…But machines can learn a lot more: inferring your age, gender, ethnicity, socio-economic status, health conditions, and beyond. Researchers have even been able to generate images of faces based on the information contained in individuals’ voice data.. soon companies may also draw conclusion about your weight [and] height[.]”.

Competition

Antitrust breach actions by Competition Commission of India (CCI)

…a couple going on.

• Tinder / Match Group v Apple: Tinder’s complaint filed to the CCI says forcing all in-app purchases through Apple’s pay system and charging high commission is anticompetitive. One interesting argument Tinder is reported to have made is that they have to pay a commission on its in-app purchases. Apple doesn’t charge developers for physical goods and services purchased through the app. It’s why Uber is exempt from paying Apple the commission – Tinder argues that Apple’s treatment of them is unfair; they are just like Uber, bringing people together, yet Apple has chosen to levy Tinder a commission. 

• CCI initiated investigation into Meta/Whatsapp, following privacy policy changes which said that users’ data will be deleted unless the new privacy policy is accepted. The investigation will also look at whether “excessive data collection” of consumers, and the sharing of the data could be anticompetitive. What I can’t really understand here is that Whatsapp messaging is understood to be end-to-end encrypted meaning Meta will not have access to the private messaging. That in itself is seen to be a big issue in India because of the potential spread of disinformation / ease of anti-government campaigns that can spread across the country without anyone being able to do anything about it. It is therefore not entirely clear to me what it means by excessive data collection, it has to be said. Any thoughts, I’d be interested…

Microsoft changes licensing terms to help small cloud providers to compete more effectively

…this happens to have followed a complained filed by OVHcloud, a French cloud service provider among others. The allegation was as follows:

• Microsoft’s licensing arrangement makes it more expensive for users to switch between cloud providers
• Microsoft software does not perform well on a non Azure [Microsoft] cloud platform

Microsoft is subject to another complaint, by NextCloud (a European company) which complains of Microsoft’s bundling of Onedrive and other Microsoft services. The EU Commission is yet to investigate the complaint. The solution Microsoft proposes tackle some of these complaints. The main one is that Microsoft software (most notably Office) can run well on other cloud services. These changes do not apply to Microsoft’s major cloud rivals, such as Amazon, Google and Alibaba.

Business

India’s conglomerate Jio ties up with WhatsApp (Meta) to enable users to browse and buy across the whole JioMart grocery catalogue without leaving WhatsApp

…if successful, WhatsApp may well become a super app (an app that deals with multiple aspects of a person’s life – eg. banking, chats, shopping etc). Jio’s Reliance Retail is largest in India, but its e-commerce presence has thus far been less significant. Whatsapp’s adoption in India is significant, being well over a third. Put them together, and you could end up with a very useful service. The challenge for WhatsApp is then to make sure that by adding the service, the app doesn’t become too cumbersome – e.g. too many ads – for the Jio tie up could only be a start for WhatsApp before it expands to other services.

Cyber security

Hackers create realistic hologram of Binance Chief Communication Officer for scamming purposes

Hackers created a hologram of Binance Chief Communications Officer which offered opportunities to list tokens on Binance over a Zoom call – this entailed an upfront fee.  Hackers created the deepfake hologram by using Binance CCO’s past appearances on TV. It was real enough to fool very tech savvy people, the CCO said.  

EV/AVs

South Korea concerned about impact of US Inflation Reduction Act

…The Act gives tax credits to EV makers but excludes EVs assembled outside the US. South Korea, home to car, battery and semiconductor makers Hyundai, Kia, Samsung, SK Hynix and LG to miss out as would other automakers of other countries such as Germany and Japan that export EVs to the US. South Korea aims to coordinate with EU and Japan to negotiate with the US administration.

Honda and LG Energy Solution to build $4.4billion EV battery plant in the US

…EV battery factories are enormously costly to build. Rising energy prices and prices of ever scarcer rare metals required to make them do not help the situation. No doubt Honda aims to qualify for the tax credit available under the Infrastructure Reduction Act – as explained above. The bans on sale of ICE cars from 2035 in Washington and California (see next) would also present a boon to the business.

BMW to produce Hydrogen cell cars too

…hydrogen-powered vehicles are ideally placed technologically to fit alongside battery-electric vehicles and complete the electric mobility picture – says BMW.

Washington and California states to ban the sale of Internal Combustion Engine (ICE) cars from 2035

…sale of used ICE cars will be allowed. Making sure that charging infrastructure is up to scratch will be key. The announcements follow that of the EU, which confirmed the ban of new ICE cars from 2035, two months or so ago.

Revel, Fermata Energy, and NineDot Energy Launch First V2G System in New York

…the G stands for Grid. Fermata Energy’s V2X bidirectional charging system can charge EVs as well discharge the stored energy in the EV batteries back to the Grid. When the demand on the Grid is high, EVs with surplus energy can supply energy back to the grid, enabling “grid resilience”.

Whatever the loss of energy entailed in charging the EVs and discharging back to the Grid, one assumes it can only improve, if it is significant. My limited understanding of how the energy system works is that, energy needs to be constantly generated; thus during sleeping hours, you end up with surplus energy that is not fully utilised. That can be used to charge EVs overnight, using cheap electricity rather than let it go to waste; if it isn’t needed then you can return that back to the Grid – making a profit in the process. Park, Plug in and Profit, is the idea.

A bit about the three companies:

Revel – provider of infrastructure and services for EVs.  

Fermata Energy – develops and owns proprietary vehicle-to-everything (V2X) software platform and bidirectional chargers turn EVs into mobile energy storage assets

NineDot Energy – currently focused on developing battery energy storage sites in the New York area.

California passes bill to protect driver data and consumers from being exposed to misleading advertising

Senate Bill 346: This bill would prohibit any images or video recordings collected through the operation of an in-vehicle camera from being used for any advertising purpose or being sold to any third party – unless the consumer affirmatively consents.

Senate Bill 1398: The bill would prohibit a manufacturer or dealer from deceptively naming or marketing autonomous features. See for example California DMV’s complaint against Tesla, over the label full self-driving description, in relation to a vehicle which is not completely autonomous.

Green Agenda

Panasonic joins the Low Carbon Patent Pledge

…alongside Hewlett Packard Enterprise (HPE), Microsoft, and Meta, and others who are pledging to make freely available their IP concerning key patented technologies that have applications in low carbon innovations including the generation, storage, and distribution of low carbon energy.

EU says certain mobile spare parts must be made available for at least five years from launch

…and  batteries should survive at least 500 full charges without deteriorating to below 83 per cent of their capacity [but this is difficult to prove]. Furthermore, software updates must be provided for 5 years after devices leave the market. Will the US [especially California] follow?

Intellectual Property

US seeks public comments to help identify potential internet and physical markets that are involved in intellectual property infringement to add to the Notorious Markets List, with a focus on online piracy and its impact on US workers

…deadline for submitting comments is 7 October.

Metaverse/NFTs

Former health secretary Matt Hancock becomes the first UK member of parliament to launch into the Metaverse

…the FT article about it, is funny not because of the content, but because of the sarcastic overtone.

Very difficult to obtain a metaverse related TM in China points out an article

…According to an article, it is difficult to obtain metaverse related trade mark rights, such as digital goods, leading to a rejection in 80% of the applications, because it is “likely to mislead consumers” – whatever this means. Furthermore, China National Intellectual Property Administration (CNIPA) explained that it requires “allowable” descriptions of underlying goods and services, which is difficult in the case of virtual goods/services because it has not yet created new subclasses for them, although there are some applications with descriptions which use terms such as “virtual worlds” and “virtual goods” which have been found to be allowable. Without being able to rely on trade mark rights for use in the metaverse, brand owners may have no option but to rely on China’s Unfair Competition law, which will be costly, the article says.

The EU IPO by the way, is seeking comments on its approach about applications that relate to the metaverse/ NFTs. Deadline for comments is 3 October. It states:

• Virtual goods are proper to Class 9 because they are treated as digital content or images. However, the term virtual goods on its own lacks clarity and precision so must be further specified by stating the content to which the virtual goods relate (e.g. downloadable virtual goods, namely, virtual clothing)

• The 12th Edition of the Nice Classification will incorporate the term downloadable digital files authenticated by non-fungible tokens in Class 9. NFTs are treated as unique digital certificates registered in a blockchain, which authenticate digital items but are distinct from those digital items [but some NFTs authenticate physical items – so what about them?]. For the Office, the term non fungible tokens on its own is not acceptable. The type of digital item authenticated by the NFT must be specified.

• Services relating to virtual goods and NFTs will be classified in line with the established principles of classification for services

Satellites/ Space

Starlink/T-Mobile alliance to enable coverage in rural parts of the US

…The significance

• Use of normal smartphone to connect to Starlink satellite; not an expensive satellite phone
• Technically significant – use of T-Mobile’s cellular 5G spectrum in this context is “like putting a cellular tower in the sky”.
• As well as enabling those in rural areas to stay connected (although quite how good the connection is expected to be is a big question mark), it could also save lives – not only of those living in rural areas but also those that have gone missing in the depths of North America. It could potentially support IoT communications.  

The technology has yet to be demonstrated.

Federal Communication Commission’s OK of Space X request to fly its satellites at a lower altitude decided with due consideration rules DC Court of Appeal

… The Communications Act of 1934 authorizes the FCC to grant radio station licenses, including for the operation of communications satellites. The Commission may modify licenses if it finds that the modification would serve the public interest among other things. The Telecommunications Act of 1996 requires the agency to facilitate the provision of broadband internet service to unserved areas.

Technical changes can interfere with signals from other satellites, so the Commission must find that the proposed modification does not present any significant interference problem. The FCC found that the changes would impose no undue interference and would serve the public interest.

Competitors had opposed:

• DISH – said changes would interfere with its GSO satellite television service. [geostationary satellite systems have to be prioritised over non-geostationary satellites, like that of SpaceX]
  • Court found that FCC’s certification of compliance to SpaceX provided assurance that there will be no harmful interference.

• Viasat/Balance Group (environmental organisation) – said environmental assessment should have taken place before deciding on SpaceX’s request but did not.
  • Court said that the parties did not have standing.

Semiconductors

Chip designer ARM sues Qualcomm, its biggest customer

…the suit is in Delaware. ARM is one of the most successful semiconductor and software design companies, mainly for CPUs (computer processors – the brains in the computer).  Its business is the licensing of intellectual property that pertain to CPUs and other chips.  ARM accuses Qualcomm of bypassing the need to take the requisite chip design licence from ARM, instead using Nuvia’s designs without ARM’s permission. ARM had said that Nuvia cannot assign its licences to Qualcomm and so Qualcomm cannot use Nuvia’s design without ARM’s blessing – yet are continuing to use Nuvia’s licensed technology. ARM terminated Nuvia’s licence.

Last year, Qualcomm bought Nuvia, a chip start-up, in the business of customizing processor core designs for data center servers, using Arm’s technology. Qualcomm says that all its acts fall within its own ARM licence.

Supply Chain

Wingtech (China) in process of building a massive factory to make PCs – thought to be to supply Apple

… Whilst Apple has recently shown moves to reduce Chinese dependency moving its manufacturing to Vietnam and India, at the same time it may be trying to reduce Taiwanese dependency too, observes Nikkei. In addition to Wingtech, Apple has procured other Chinese companies to make its products for them.  

Note:

• Making MacBooks is technically demanding, so much so that it was only thought to be the mainstay of Taiwanese Quanta and Foxconn. Even the Taiwanese Compal and Wistron are thought maybe not to be up to scratch despite their trusted supplier status to leading PC brands like HP, Dell and Lenovo.

• The selection of Wingtech to make the new M2 powered MacBook Air under the circumstances is a massive endorsement.

• already makes PCs for Samsung, Lenovo and Asustek, according to some sources
  • also makes phones for successful Chinese makers Oppo, Vivo and Xiaomi
  • bought chipmaker Nexperia from NXP in 2019 and acquired the U.K.’s largest chip plant (Newport Wafer Fab – though currently under investigation on national security grounds) through Nexperia in 2021

We will hear more about Wingtech in the future no doubt.     

Telecoms

UK Government puts on further pressure on telco and network companies by proposing to add new regulation on ensuring security

…pressure which telco companies don’t need. The infrastructure cost entailed in upgrading to 5G is so significant, telco companies are lobbying big tech companies and streaming companies who stand to gain from the 5G offering, to contribute. In the UK there are four carriers, but business is not easy, so much so that the smaller of the two – Vodafone and Three are in merger talks. That is not to say what the government is proposing is wrong. There has been a spike in hacking and security breach incidents. Increasing geopolitical tensions only adds fuel to this fire. The importance of connectivity in our lives also makes the problem acute – one just has to imagine the mayhem that might ensue if the autonomous vehicle network were to be hacked. I’ve heard it described that data scientists and information security analysts will never be out of the job in the future.

The rules follow the Telecommunications (Security) Act 2021 [which amends the Communications Act 2003] to give the government new powers to make regulations and issue codes of practice to provide guidance on how to comply with the new regulations. In essence, it introduces new duties on providers of public electronic communications networks and services (‘providers’) to identify and reduce the risk of security compromises, and prepare for the possibility of their occurrence [meaning they would have to anticipate security attacks] and places duties on providers to prevent, remedy or mitigate any adverse effects of security compromises.

It used to be the case that providers could decide on the appropriate standard, but that meant that the standards could often be less than adequate.

It would be interesting to see if the new regulatory burdens would impact on the adoption of Open RAN – something the UK Government has funded the development of, and is keen for us to adopt to encourage competitiveness. However because it will have open interfaces, it presents hackers with opportunity for entry.

Short note on Open RAN

The current wireless connections are provided by a hardware/software integrated platform (there are three main ones in the world – Nokia, Ericsson and Huawei – and maybe ZTE and Samsung too) that provide a telecommunication service end to end. In an Open RAN configuration, technologies are not integrated – software has open interfaces and can be sourced by variety of suppliers, and are hardware neutral – so hardware too can be sourced by a variety of suppliers, and in fact commercial off the shelf parts for PCs can be used. So if you need to upgrade a piece of hardware or software, you have a number of vendors you can source from, which encourages competitiveness.   

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.