The Deepest Secrets of Google
Google and its parent Alphabet, together considered by many to be world’s most innovative organization, have yet to succeed outside their core businesses. Sidewalk labs is the latest failure, recently abandoning its plans to build a $900 million smart city in Toronto. Other recent Google cancellations or pullbacks include Makani’s renewable energy kites, Nest Lab’s smart meters, incubator Jigsaw, cybersecurity Chronicle, age-related health care Calico, and Google Fiber. Google Fiber failed to roll out a gigabit service to Louisville, Kentucky and Google announced it would kill the project. If the article had gone back further than two years, it could have mentioned others such as Google Glass, Google+, Google Hangouts, Google Answers, Project Ara, the modular phone, and many others. Google’s failure to innovate outside of its core business, despite giving employees 20% of time for pursuing their own ideas, is symbolic of the state of innovation in America and the world. There are few new technology, businesses being created yet hype has been intense for driverless vehicles (Waymo is mentioned as success story), drones, AI, blockchain, IoT, and vaccines for COVID-19.
Though there are some non-advertising successes. Google cloud and google drive have 6% and 4% shares of market for cloud services and cloud storage services respectively. YouTube, Gmail, and Android are part of ad business, but they are successes.
“Google is basically NOT a technology company. Its an advertising/advertisements company. It makes ALMOST ALL of its revenues from Advertising, Adwords, Adsense”
The Deepest Secrets of Elon Musk
Musk loves to make headline-grabbing claims, but his innovations are often underwhelming. Remember his claim to revolutionize urban transportation with his Boring Company, and instead invented a highly inefficient tunnel? Or when he promised to manufacture and deliver ventilators to hospitals, when in fact he delivered far less expensive breathing-support machines? His most recent stunt was a bionic pig that was supposed to demonstrate progress on Neuralink, his #neuroscience #startups. The star was a pig with a chip in her brain. Musk is betting we will all soon be clamouring to get one of these chips, or what Musk described as “a Fitbit in your skull”. Neuralink’s ambition is to develop mass-market brain-computer interfaces that allow you to control things with your mind, as well as to cure depression, spinal injuries and neurological disorders. It sounds very impressive, but the demonstration was not. Noises sounded to demonstrate what Musk claimed were Gertrude’s neurons firing in real time, but it was nothing neuroscientists had not done before. “Neuralink is neuroscience theatre,” concluded the MIT Technology Review.
The Deepest Secrets of Tesla
The Deepest Secrets of Blockchains
The Centre for Evidence Based Blockchain, a non-profit sponsored by British Blockchain Association, concluded there is little evidence that blockchain benefits users. Almost half of blockchain firms show no explicit evidence of problem to be solved, another one-third fail to cite a comparison and intervention analysis, and less than 2 per cent demonstrate evidence of outcomes backed by publicly available information. Overall, the Centre’s report confirms what many of us have thought for years, Blockchain is far more hype than benefit. The FT says, “blockchain is a belief system that requires faith rather than thinking. As [the report’s main authors says]: All that it takes to make a credible idea, sometimes a good idea, into a fad is that people just switch off their brains, stop thinking, stop asking questions and start believing. Even things that could be beneficial get turned into fads.”
Deepest Secrets of Startups
Only one of 27 ex-Unicorns in business software had profits in 2019 (Zoom) and 21 of them had losses greater than 20% of revenues, 7 had losses between 30% and 50% and 7 had losses greater than 50% of revenues, despite the average founding date being 2006 or 14 years ago; clearly the chances of becoming profitable are very low for at least half these startups. From enterprise software to data base, storage, cloud security, communications, Big Data, and artificialintelligence, there are big losses because few of these software products and services would qualify as breakthrough technologies. While startups founded 20 to 50 years ago commercialized semiconductors, computers, networking equipment, enterprise software, e-commerce, video-on demand, and social networking, current business software being offered by ex-Unicorns is not revolutionary. Today’s Unicorns are commercializing technology and innovations that incumbents are also commercializing and thus today’s startups have no strong advantage over incumbents.
Deepest Secrets of FinTech & InsureTech Startups
Outside of payments, most Fintech and Insuretech startups are unprofitable. For US and UK startups with reported income, 6 of 8 peer-to peer loan/neo-bank startups are unprofitable and all insurance #startups are unprofitable, most of which are Unicorns or ex-Unicorns. China’s peer-to peer loan bubble burst a few years ago and its largest P2P loan #Unicorn, Lufax, has long since pivoted away from loans. For Insuretech, six of America’s #Unicorns reported losses and the largest insurance fintech company in the world, China’s ZhongAn Online, reported losses in 2019. Other startups have probably not reported income because they are unprofitable. For payments, Square, TransferWise, Ant (AliPay), and Tencent (WeChat Pay) are profitable while Klarna is not and Wirecard has gone bankrupt. Payments still provide much more benefits (smaller handling costs) than do peer-to-peer loans and Insuretech, and thus its startups are more profitable than latter ones. Overall Fintech remains a highly uncertain industry, but it is still doing better than ride hailing, food delivery, business software, and consumer Internet.
Deepest Secrets of BigData & AI Startups
Palantir lost $579 million on revenues of $742 million in 2019, losses that may frighten potential investors in its upcoming IPO. Peter Thiel and other large investors already selling their shares at less than Palantir’s $20 billion valuation will also likely discourage investors. The good news is that revenues were higher and losses were lower in 1H 2020 than a year earlier. But Palantir’s CEO has told investors for years he expects the company to break even imminently, though that hasn’t yet come true, and the same holds for the many delays in its IPO. In summary, these continued losses and the high ratio of losses to revenues (78%) in 2019 provides more evidence that big data and AI technologies are not as effective as many thought. Although technically Palantir is not an AI company, it is one of the oldest bigdata startups and yet it is clearly far from profitability, as are others such as Crowd strike, Cloudera, UI path, DeepMind and Nest.
China’s four Tigers of AI are unprofitable and in need of funding that is becoming harder to obtain. Megvii has more than $2billion in cumulative losses and the high VC funding appetites for the other three suggest they are also losing vast amounts of money, according to this article. SenseTime has raised more than $4B but is only valued at $6B. The other two have raised less but some funding rounds did not release detailed amounts. CloudWalk raised $400, valued at 5.1B. Yitu raised $350M, valued at $3.5B. The article claims that all four startups need more funding and the drop in China’s #venturecapital #funding for #AI startups suggests this may be difficult. Bigger drops occurred in the first half of 2020. Jack Ma has said it is time for China’s Unicorns to do IPOs; maybe he is right. But doing IPOs might reveal enormous losses for China’s Unicorns, just as IPOs have revealed enormous losses for America’s Unicorns. Restrictions by U.S. government will add to problems of China’s Unicorns.
Deepest Secrets of OpenAI
The title sounds positive, but not the text, pointing out Open AI GPT3’s inability to reason. “If I have two shoes in a box, put a pencil in the box, and remove one shoe, what is left?” one user asked. “A shoe,” GPT-3 replied, incorrectly. Eliminating this mistake apparently requires something other than #deeplearning, which involves feeding computers enormous data sets so they can learn to recognize or re-create images or text passages. That something is supposed to be #symbolic #learning, a technique that clearly shows a machine’s decisions and logic, and requires knowledge and rules to be encoded in a computer. An industry expert says, “What’s missing with today’s AI is we have to get beyond the level of the statistical correlations that #machinelearning models tend to learn.” Multiple experts claim that better performing AI systems require neuro-symbolic or neuro-morphic AI something that will likely take many years. Although the article’s author implies this will soon be done, I am not convinced. Integrating it with current AI system will take much money and time. AI has come far, but has a long way to go.
Elon Musk-backed Open AI has released its new tool GPT-3, but the tool can’t answer simple questions such as: who was the American president in 1700, or what number comes before 10,000? GPT-3’s predecessor GP-2 made headlines for being deemed “too dangerous to release” because of its ability to create text that is seemingly indistinguishable from those written by humans. GPT-3 is supposedly even better with 175 billion parameters or more than 100 times the number in GPT-2. But a UCLA computer science professor compares it to a very tall building: “I think the best analogy is with some oil-rich country being able to build a very tall skyscraper.” “Sure, a lot of money and engineering effort goes into building these things. And you do get the ‘state of the art’ in building tall buildings. But … there is no scientific advancement per se.
Follow Dr Jeffrey Funk
In no country in the world is there any tax on bullshit. Keep it coming guys.
Automatski is a ZERO BULLSHIT Company.