Tech Brew Ride Home - Fri. 08/07 – Twitter Blocking Substack?

Episode Date: April 7, 2023

Shenanigans from some Elon companies. Twitter looks like it’s blocking Substack. Not just links to Substack but even mentioning the name, the word Substack, in some cases. Oh, and Tesla employees ap...parently liked to view videos from your car and share them with each other internally. Last quarter was epically bad when it came to venture investing. And, of course, the Weekend Longreads Suggestions. Links: Twitter now disables likes, replies, and retweets if a tweet has Substack links (The Verge) Special Report: Tesla workers shared sensitive images recorded by customer cars (Reuters) A third of organizations admit to covering up data breaches (VentureBeat) Global VC Funding Falls Dramatically Across All Stages In Rocky Q1, Despite Massive OpenAI And Stripe Deals (CrunchBase News) Weekend Longreads Suggestions: Behind the curtain: what it feels like to work in AI right now (Democratizing Automation) Think of language models like ChatGPT as a “calculator for words” (Simon Willison's Blog) Anthropic’s $5B, 4-year plan to take on OpenAI (TechCrunch) Blockbuster pushed HBO to start investing in original content, cable giant’s ex-chief says (CNBC) Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript
Discussion (0)
Starting point is 00:00:04 Welcome to the Tech meme right home for Friday, April 7th, 2020. I'm Brian McCullough today. Shenanigans from some Elon companies. Twitter looks like it's blocking substack, not just links to substack, but even mentioning the name, the word, substack in some cases. Oh, and Tesla employees apparently like to view videos from your car and share them with each other internally. Last quarter was especially bad when it came to venture investing, and of course, the weekend long read suggestions. Here's what you miss today in the world of tech. shenanigans. Substack is telling its writers Twitter, quote, unexpectedly restricted access to embedding tweets from Substack on Twitter. Not great for Substack writers. Twitter is maybe the main platform most Substack writers used to promote their work. But wait, there's more.
Starting point is 00:00:55 Twitter seems to have disabled likes and retweets for tweets with Substack links as well. So even if you're not a substack writer, if you tweet a link to a Substack post, your tweet might be hobbled. Also, some people are reporting that even tweeting the word Substack could affect your tweets. I guess Elon didn't take kindly to the launch of that Twitter clone by Substack, quoting the verge. The issue could cause problems for writers who want to talk about what's going on with Twitter in their newsletters or about things that are happening on the platform. While screenshots of tweets could work in some cases, they're less trustworthy because they don't provide a direct link to the source. screenshots also won't help you if you're trying to, say, embed a video that someone posted on Twitter,
Starting point is 00:01:39 and Twitter seems to be at least somewhat interested in becoming a video platform given that several blue perks relate to making the video uploading experience better. There are several reports that tweets including links with substack in the name are being restricted with an accompanying message that says some actions on this tweet have been disabled by Twitter. Although not everyone, including the verge at the time of this update, is seeing this behavior. Those affected reportedly can't like the offending tweet, reply to it, or retweet it. This could be related to the Twitter API changes. Various companies have been notifying users that they have to cut off or paywall certain features that interacted with Twitter. And many people who have run bots on the platform have been posting about how they can no
Starting point is 00:02:22 longer post like they used to. The API transition has been bumpy, even for those that appear to be in Twitter's good graces. Earlier this week, WordPress's API access was suspended, making it so users couldn't auto-share posts to the platform. The company was eventually able to get it restored and says that it'll be, quote, working with Twitter directly to ensure this service keeps running without interruption, end quote. There are several tools that integrate Twitter that do still work. Embeds still work in Ghost, a blogging platform similar to Substack, as well as in the Verges' content management system. However, if those tools rely on API access to work,
Starting point is 00:02:56 it's possible there could be problems ahead as Twitter continues to deprecate access to it, end quote. More shenanigans. Roiders is reporting that Tesla staff privately shared videos from customers' car cameras from 2019 to 2022 on an internal messaging system, including a video of a child being hit by a car. Quote, between 2019 and 2022 groups of Tesla employees privately shared via an internal messaging system, sometimes highly invasive videos and images recorded by customers' car cameras according to interviews by Reuters with nine former employees. Some of the recordings caught Tesla customers in embarrassing situations. One ex-employee described a video of a man approaching a vehicle completely naked. Also shared crashes and road rage incidents. One crash video in
Starting point is 00:03:48 2021 showed a Tesla driving at high speed in a residential area hitting a child riding a bike according to another ex-employee. The child flew in one direction, the bike in another. The video spread around a Tesla office in San Mateo, California, via private one-on-one chats like, quote, wildfire, the ex-employee said. One ex-employee also said that some recordings appeared to have been made when cars were parked and turned off. Several years ago, Tesla would receive video recordings from its vehicles even when they were off if owners gave consent. It has since stopped doing so. We could see inside people's garages and their private properties, said another former employee. Let's say that a Tesla customer had something in their garage that was disliked.
Starting point is 00:04:28 You know, people would post those kinds of things, end quote. About three years ago, some employees stumbled upon and shared a video of a unique submersible vehicle parked inside a garage, according to two people who viewed it. Nicknamed Wet Nellie, the White Lotus Espree Sub, had been featured in the 1977 James Bond film, The Spy Who Loved Me. The vehicle's owner, Tesla Chief Executive Elon Musk, who had bought it for about $968,000 at an auction in 2013. It is not clear whether Musk was aware of the video or that it had been shared, end quote.
Starting point is 00:05:07 A Bit Defender survey of more than 400 IT security professionals has revealed that around 42% of them were told to cover up data breaches that should have been reported, and 29.9% of them admitted to covering up those breaches. Quote, this research highlights that an alarming number of organizations are willing to ignore their obligations to report data breaches to regulators and customers in an NNN. attempt to avoid legal and financial penalties. The research comes less than a year after the FTC convicted former Uber CSO Joseph Sullivan for attempting to cover up a hack of Uber in 2016. The case highlighted that lying about data breaches is a serious criminal offense in many jurisdictions. So why are so many tech leaders pressuring their staff to bury data breaches? The answer is that the cyber threat landscape is becoming more and more demanding with 52% of organizations experiencing a data breach within the past 12 months. The five threats that respondents reported they are most concerned about
Starting point is 00:06:02 are software vulnerabilities and zero days at 53.9 percent. Fishing and social engineering at 52.2 percent. Supply chain attacks, 49 percent. Ransomware, 48 and a half percent. And insider threats, 36 and a half percent. Worldwide organizations are under tremendous pressure to contend with evolving threats such as ransomware, zero-day vulnerabilities and espionage, while struggling with the complexities of extending security coverage across environments and an ongoing skill shortage, said Andre Flerescu, deputy general manager and senior vice president of products at Bit Defender Business Solutions Group, end quote. And this is hugely notable for all startups out there. Last quarter, Q1 of 2023, was a terrible quarter for investments.
Starting point is 00:06:52 Global VC funding fell 53% year-on-year from $162 billion in Q1 of 2022 to just $76 billion in Q1 of 20203. But it's even worse than that, because if you take out open AIs reported $10 billion, and stripes reported $6.5 billion raises, you're only looking at around $60 billion, and of that, only $39.2 billion went to late-stage startups. That's notable because late-stage rounds are usually the largest in aggregate. So this means that if you're a company looking for a series B or beyond, wallets do no seem to be open. right now. Quoting Crunch Base. Every funding stage last quarter was down 44 to 54% year over year, a clear signal that the slowdown is not confined to late stage funding. Investors across each stage
Starting point is 00:07:43 scaled back as they took time to assess new investment opportunities while guiding existing portfolio companies. In the first quarter of 2023, seed funding totaled $6.9 billion down 44% year over year. They signal that even at the earliest funding stages investors are pulling back. That's significant because seed funding was by far the least impact. funding stage through the 2022 reset. In the first half of 2022, seed funding increased each quarter based on a year-over-year comparison. It was only in the fourth quarter of 2022 that global seed funding amounts tended down year-over-year by more than 25%. Despite the slowdown investors and private companies still hold record amounts of dry powder with around $580 billion as of the end of
Starting point is 00:08:25 2020, according to an estimate by James Erfetti of Lightspeed Venture Partners. This is in line with the amount of dry powder available in 2021, but investors that year were plowing the money rapidly into startups. Early-stage funding totaled $25.6 billion in Q1 down 54% year-over-year. The year-over-year pullback in early-stage quarterly funding in 2022 started in the third quarter and has continued to decline since. Late-stage funding totaled $43 billion, a dramatic fall from $93 billion in Q1, but up from $34 billion in Q4. The billions of dollars raised by Open AI and Stripe made up 22% of all venture capital raised this past quarter and 38% of late stage financings, end quote. Time for the weekend long-reach suggestions. First up, we've all been dealing with wave after wave of
Starting point is 00:09:21 news about this new AI stuff, suddenly impacting seemingly every corner of the tech industry. But what about if you've been in AI for a while, like a decade or more? What if this has always been your corner of the tech industry? What has this moment been like for you? Well, here you go. From the democratizing automation substack, quote, Every single person I know working in AI these days in both the academy and industry has been sparked by the chat GPT moment, the first iPhone moment of AI. Working in this environment is extremely straining for a plethora of reasons, burnout, ambition, noise, influencers, financial upside, ethical worries, and more. The chat GPT spark has caused career changes, projects to be abandoned, and tons of people
Starting point is 00:10:06 to try and start new companies in the area. The entire industry has been collectively shaken up. It added a ton of energy into the system. We now have model and product announcements on almost a daily basis. Talking to a professor friend in NLP, it's to the point where all sorts of established researchers are ready to jump ship and join or build companies. This is not something that happens every day. Getting academics to stop wanting to do research is a hilarious accomplishment. Everything just feels so frothy. Graduate students. are competing with venture-backed companies. From a high-level technologist's perspective, it is awesome.
Starting point is 00:10:42 From an engineer-on-the-grounds perspective, it leaves some stability and naps to be desired. Seeing all of the noise makes it hard to keep one's head on straight and actually do the work. It seems like everyone is simultaneously extremely motivated and extremely close to burning out. Given the density of people in all the project spaces of generative AI or chatbots, there is a serious be-the-first or be-the-best syndrome with a third axis of success. being openness. This keeps you on your toes to say the least. In the end, these pressures shape the products people are building away from a thoroughness of engineering and documentation. Clickiness is the driving trend in the last few months, which has such a sour flavor, end quote.
Starting point is 00:11:24 Then Simon Willison has a blog post that made an analogy about this new AI stuff that I thought was useful. Quote, one of the most pervasive mistakes I see people using with large language model tools like chat GPT is trying to use them as a search engine. I like to think of language models like chat GPT as a calculator for words. This is reflected in their name. A language model implies that they are tools for working with language. That's what they've been trained to do, and it's language manipulation where they truly excel. I want them to work with specific facts, paste those into the language model as part of your original prompt, end quote. He then goes on to give you basic suggestions for how best to prompt the bot, if you think of it,
Starting point is 00:12:06 as a calculator for words. Then it's amazing how quickly major players have already seemingly sprung up for this new era in this new space. If Open AI is already the 800-pound gorilla is Anthropic, going to be the, I don't know, Pepsi to their Coke, quoting TechCrunch. AI research startup Anthropic aims to raise as much as $5 billion over the next two years to take on rival OpenAI and enter over a dozen major industries, according to company documents obtained by TechCrunch.
Starting point is 00:12:37 A pitch deck for Anthropic Series C fundraising round discloses these and other long-term goals for the company, which was founded in 2020 by former Open AI researchers. In the deck, Anthropic says that it plans to build a frontier model tentatively called Claude Next, 10 times more capable than today's most powerful AI, but that this will require a billion dollars in spending over the next 18 months. The information reported in early March that Anthropic was seeking to raise $300 million at a $4.1 billion valuation, bringing its total raise to $1.3 billion. The deck confirms that target number, though only half was raised at the time of the
Starting point is 00:13:14 documents creation from a confidential investor. Anthropic describes the frontier model as a next-gen algorithm for AI self-teaching, making reference to an AI training technique it developed called constitutional AI. At a high level, constitutional AI seeks to provide a way to align AI with human intentions, letting systems respond to questions and perform tasks using a simple set of guiding principles. Anthropic estimates that its frontier model will require, on the order of 10 to the 25th flops or floating point operations, several orders of magnitude larger than even the biggest models today. Of course, how this translates to computation time depends on the speed and scale of the system doing the computation. Anthropic implies in the deck it relies on
Starting point is 00:14:00 clusters with tens of thousands of GPUs, end quote. And finally today, another classic story of corporate competition from another era. We remember, most of us, how Netflix's first big battle was to supplant Blockbuster. But former Time Warnerhead Jeff Bukes has revealed something I had never thought of. HBO, he claims, became the HBO we grew to love only because of competition not from Netflix, but from Blockbuster. Those of you too young to know this, HBO used to only be movies, well, movies and boxing and stand-up specials, but until the 1990s, they didn't do original series. They didn't do original
Starting point is 00:14:43 programming at all. They only got into that game because, again, of Blockbuster, quoting from CNBC. The first Blockbuster opened in 1985. Through the late 1980s, Blockbuster Undercut HBO's Business Proposition. Consumers now had another way to watch uncut commercial-free movies in the house and now they could do it on demand. HBO only broadcast one movie at a time. Blockbuster offered thousands of movies. Movie studios viewed the rental market as a booming industry and offered Blockbuster its films six months ahead of HBO's window.
Starting point is 00:15:14 Movies were our number one selling point, said former HBO and Time Warner Chief Executive Officer Jeff Buick's, who spoke to CNBC as part of a digital documentary on the history of HBO. We're sitting at HBO saying we're screwed. Our number one reason why people want to subscribe to HBO is now being taken by Blockbuster. What are we going to do? HBO at the time, led by Michael Fuchs, decided the answer was original programming. With only a few million dollars to spend, Fuchs placed two small bets on two original comedy series, Dream On, which premiered in 1990, and the Larry Sanders show, which arrived in 1992, end quote.
Starting point is 00:15:49 It was Bukes, who was CEO when Sex in the City, Oz and the Sopranos turned HBO into the HBO we know and still mostly love. All right, a metric ton of show news to share with you right here. First up, we do have a bonus episode for you this weekend. Chris and I spoke with the great David Auerbach about his assertion that major tech platforms are so big these days that no one, maybe even especially the people working inside these platforms, have control of them. Meaning they don't even know how their big monsters work. The monster is out of the lab. This is, shall we say, a provocative idea in this age of AI, so look for that. Then, very important news for you, that long-discussed trip to Ireland is happening next week.
Starting point is 00:16:47 No, I will not be doing the robot voice next week. Instead, Jackson Bird, the former host of our former sister show. Cool stuff right home will be filling in for me all week. And because he will be doing so, the show will post slightly later all next week. We're aiming for the window between 3 p.m. and 4 p.m. Eastern, so noon and 1 p.m. Pacific. This is because Jack will need a bit more time than I usually need to get everything together. And if we aim to post in that window, it'll also be past 8 p.m. Dublin time so I can be online to assist him. I'll try to feed Jack what stories I think he should go with as best I can.
Starting point is 00:17:23 But never fear. Jack did a daily show for us for the better part of three years. He's a pro. He'll be great. So the show will go on all next week. The next time you will hear from me will be April 17th. Talk to you then. Behave yourself in my absence.

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.