Peter's Quasi-Daily Blah-g

• Tesla driver Tim Heckman drove 6,392 miles primarily using Autopilot and Full Self-Driving.
• He said Autopilot has gotten “worse” over the years and FSD was “exceptionally poor outside of California.”
• But, the Tesla driver said the software has been a “lifesaver” when it comes to long road trips.

A Tesla owner took a 6,392-mile road trip using primarily Tesla Autopilot and Full Self-Driving (FSD) — and said that while the software was a “lifesaver,” there were some hiccups along the way.

In December, Tim Heckman drove a Model S Plaid from Los Angeles to Pennsylvania and back, using the autonomous software for 99% of the journey, an experience he documented on Twitter.

Heckman, a site reliability engineer, told Insider that while the autonomous software proved helpful during his journey, it also made for a “stressful drive” at times, detailing incidents where the technology phantom braked and struggled to obey the speed limit, proper following distances, or stay in its lane. 

The pros and cons of autonomous driving

While Autopilot is a driver assist software that is built into all Teslas and designed for driving on highways, FSD is a beta add-on that can operate in urban settings and is designed to change lanes, recognize stop signs and lights, as well as park.

Tesla CEO Elon Musk has said that the software will eventually be able to operate entirely on its own and will be safer than human drivers, but the beta program still requires a licensed driver to monitor it at all times.

Heckman told Insider the software sometimes registered cars on the screen that weren’t there or had difficulty identifying lane markings when there was salt on the road.

“It’s kind of like driving with a 15 or 16-year-old driver sometimes,” Heckman said of using FSD in city streets outside of California. “There’s weird jerky maneuvers. It’ll stop or get into a turn lane too early. In a way, there’s just a general lack of awareness about the environment.”

On the other hand, the Tesla owner said Autopilot was a “lifesaver” on highways, adding that while he had to disengage the FSD software on numerous occasions, Autopilot was only disengaged once when a car in front of him on the highway slammed on his brakes.

“It can be a huge cognitive relief. Long trips can take a mental toll,” Heckman said, noting that he’s used Autopilot on previous road trips and discovered he could drive further without getting tired.

The software has also helped him avoid collisions on the highway in the past.

“I realize I sometimes tune out when I’m driving,” Heckman said “This [software] can augment that, but I know if I do tune out, at least I know the vehicle is backing me up.”  

Getting worse, not better

In his Twitter thread about the experience, Heckman wrote Autopilot was “worse” than when he bought his first Tesla in 2019 and FSD was “exceptionally poor outside of California.”

Hedges & Company, a digital marketing firm for automakers, found in an analysis of over 175 million car owners in 2019 that the majority of Tesla owners live in California — meaning the AI software could have more opportunities to learn from California roads.

Ultimately, Heckman said he couldn’t see himself buying a non-Tesla electric-car — at least not until the charging networks could catch up to Tesla — but he wishes the automaker would rely on LiDAR, radar sensors that can help vehicles detect nearby objects.

Musk has spoken out against the expensive hardware in the past and reportedly demanded cameras over radar because he wants the autonomous software to operate like human eyes. The car company stopped installing LiDAR in its cars in 2021.

Heckman isn’t the first person to detail issues with Tesla’s Autopilot or FSD add-on. Many FSD testers have posted videos showing bugs in the software. The National Highway Traffic Safety Administration is investigating Autopilot and its potential connection to several accidents.

“At the end of the day, I think this stuff has tremendous potential,” Heckman wrote on Twitter. “But at this point there needs to be focus and good execution, while not causing regressions in the experience especially on features that impact your safety and the safety of others on the road.”

Do you drive a Tesla or have some insight to share? Reach out to the reporter from a non-work email at gkay@insider.com

Not all coffees are created equal. You already know this if you’ve been brewing your own for more than a few cups’ time: Coffee beans vary wildly in taste, strength, and even caffeine content. But even how you brew your coffee changes its strength.

I don’t just mean the flavor. If you’re drinking coffee for the caffeine boost, your brewing method can affect how much of a jolt each serving will give you. You might thing an espresso shot is the best way to get moving in the morning, but an espresso—even a full or double shot from a specialty coffee shop—actually doesn’t have the most amount of caffeine per cup—at least according to tests conducted by popular Coffee Youtuber James Hoffmann. Hoffmann is the author of The World Atlas of Coffee, runs Square Mile Coffee Roasters, and won the 2007 Barista championship, so you can count on his brewing expertise.

Hoffmann got his hands on a $2,500 Lighttells caffeine analyzer and went to town with it. This is a simple (but pricey) piece of tech reveals the caffeine content of a brew by scanning a couple of drops of coffee.

The caffeine content in a cup of coffee is measured in milligrams (mg). According to the FDA, 400 mg/day is the recommended daily intake for adults (but of course, how much caffeine you actually “need” to feel the effects will vary from person to person). Unfortunately, there isn’t exactly a wealth of data out there of coffee’s wildly varying caffeine content, but if you Google, you’ll find that on average, a cup of coffee has 63-125 mg of caffeine in it.

What coffee brewing method delivers the most caffeine?

Hoffmann tested different kinds of coffees, ranging from instant coffee, espressos, and brewed coffees like pour-overs. And what he found was very interesting. It turns out if you want a big kick of caffeine to get you through a couple of hours of intense work, the brewing method that will serve you best is a pour over. According to hHoffman’s tests, 18 grams of his test sample of coffee beans (around 2 to 2-1/2 tablespoons of ground coffee), brewed as an espresso, had 110 mg of caffeine. But brewed as a pour over, they delivered 180 mg of caffeine. That’s close to a 50% jump!

That means if you need the biggest boost, you should skip that double-shot latte from Starbucks and order a pour over. (Or better yet, start brewing pour overs at home.)

Another big takeaway from Hoffmann’s study—if an unsurprising one—is that instant coffee is, on average, quite weak when it comes to getting you caffeinated. A full cup made with 2 grams of instant coffee (the recommended serving size) will only get you around 80 mg of caffeine.

Check out Hoffmann’s complete 20-minute video for more fascinating caffeine data.

[James Hoffmann]

Long a staple technology of electronic book readers, electronic paper is finally starting to expand its reach, finding its way into tablets, watches, and even cars, with a recent collaboration at CES between E Ink and BMW. But CES also saw the reveal of an even stranger device based on electronic paper: a wearable device called the Sol Reader that’s an e-reader you can wear on your face.

Although certain technical limitations prevent electronic paper from being used as a universal replacement for LCDs and OLEDs—you definitely don’t want an E Ink TV hanging in your family room—its technical advantages, including low-power draw and a reflective screen that’s easier on the eyes, make it ideal for everything from e-readers to e-notes. Where we didn’t expect electronic paper to show up, even the new color varieties, was in a pair of goggles.

So far, the goggles form factor has been reserved mostly for either virtual reality, augmented reality, or wearable devices that mix both, leveraging two screens to immerse viewers in 3D worlds or experiences. There’s little doubt that wearables are the future of the devices we still carry stuffed away in our pockets, but the Sol Reader isn’t looking to replace smartphones. Instead, it’s targeting e-readers.

Details about the device on its official website are slim, but Brad Lynch, of YouTube’s SadlyItsBradley channel, managed to get some hands-on time with a prototype of the Sol Reader at CES last week, and shared their impressions in a video. Weighing in at less than 100 grams, the goggles look and feel like a lightweight pair of VR goggles (you can’t see the world around you while wearing them) but instead feature a pair of electronic paper screens behind a set of pancake lenses.

The Sol Reader doesn’t present users with 3D interactive virtual worlds, but instead allows them to read. You’ll see what looks like a book page floating in a dark void in front of you, or above you when using the goggles while laying down, which is apparently the ideal use case, as the Sol Reader doesn’t feature a head strap. Page turning is accomplished through a wireless handheld remote, while battery life is promised to be close to 30 hours, thanks to the use of e-paper displays.

Lynch does give a quick peek at the screens inside the device in the video, which unfortunately look disappointingly low-res, but points out the team behind it has been talking to E Ink about securing a solution with more resolution. That will be critical, as the Sol Reader is expected to sell for a hefty $350 sometime this year. If the reading experience doesn’t come anywhere close to what a Kindle provides, it’s hard to imagine anyone spending that much money on the wearable when the Meta Quest 2 is just $50 more and has plenty of e-book reading apps available for it.

Seattle’s public school district is suing Facebook, Instagram, TikTok, YouTube, Snapchat, and their parent companies. The lawsuit, filed on Friday in a U.S. District Court, alleges that these social media sites have been a primary factor in a “youth mental health crisis,” and that these platforms have knowingly exploited, manipulated, and targeted young people for profit at the expense of their mental heath.

The district argues in its 91-page complaint that tech giants have intentionally engineered addicting platforms, cashed in on the vulnerability of still-developing brains, and algorithmically suggested harmful content to young users.

Ultimately, the school district is blaming these social media companies for the increase in mental health and behavioral issues that teens are showing up to classrooms with, which has rendered the task of educating more difficult, according to the suit. District officials point to a 30% increase in self-reported feelings of sadness and hopelessness among the student body, as well as a rise in student suicide plans and attempts between 2010 and 2018.

In an effort to manage those challenges, the school district says it has had to take expensive actions like hiring more mental health counselors, creating curriculum surrounding social media and mental health, adjusting and enforcing school policies surrounding social media use, and increasing disciplinary resources. However, even all of these changes haven’t been enough to manage.

“Plaintiff cannot keep up with the increased need for mental health services because of the youth mental health crisis,” the lawsuit claims. So, the Seattle schools are seeking accountability for social media platforms and meaningful change in how these companies operate, along with damages and compensation.

In past, similar cases, tech companies have used Section 230 of the Communications Decency Act as a legal shield. Under the law, digital publishers are not responsible for third-party content posted on their platforms (i.e. Meta is not liable for anything its users post on Instagram and Facebook). However, the Seattle case aims to get around this fundamental protection by targeting the design of social media sites—not their content. The school district is claiming the increasing incentives to spend more and more time scrolling and the algorithms that dictate what users see causes harm too—not just what’s in the posts.

“Defendants have maximized the time users—particularly youth—spend on their platforms by purposely designing, refining, and operating them to exploit the neurophysiology of the brain’s reward systems to keep users coming back, coming back frequently, and staying on the respective platforms for as long as possible,” says the complaint.

Some psychology research, along with both internal and external reports on social media company practices seem to support many of the new lawsuit’s claims. Studies have shown, for instance, that social media use and increased smartphone use may be linked to sleep depravation and accompanying depression. A Pew 2022 analysis found that more than half of teenagers surveyed would have a hard, or very hard, time giving up social media. Meta’s own internal research suggested that Instagram is toxic to some teen users, particularly girls, as it cultivates and amplifies body image issues. And Facebook has known for years that its algorithms boost time spent on its site to users’ detriment.

However, it’s very difficult to establish a direct link between increased social media use and worsened mental health because there are so many variables involved in mental health. And many experts dispute the use of the term “addiction” as applied to social media platforms altogether.

This isn’t the first attempt to sue social media companies for alleged mental health or youth harms in the U.S.. However past suits have mostly focused on individual cases. For instance, the mother of a 10-year old who died in 2021 sued ByteDance over allegations that a TikTok challenge caused her child’s death. And, in April, the mother of a Wisconsin 17-year old who died by suicide sued Meta and Snapchat for “knowingly and purposely” creating harmful and addicting products. The FTC has forced Fortnite to change its interface design so as to be less deceptive (and fined Epic Games half a billion dollars).

California legislators even tried to pass a bill banning addictive social media and explicitly making tech companies liable for every resulting violation involving children. The bill failed, but more than 30 states currently have some sort of proposed or pending legislation aimed at regulating social media.

Gizmodo reached out to Meta (Instagram and Facebook’s parent company), Alphabet (Google and Youtube’s parent company), TikTok (owned by ByteDance Inc.), and Snapchat (owned by Snap Inc.) for comment.

“We want teens to be safe online,” wrote Meta’s head of global safety, Antigone Davis, in a statement emailed to Gizmodo citing tools the company has developed “to support teen and families,” like age verifications, parental controls, and notifications encouraging breaks. Further, “we don’t allow content that promotes suicide, self-harm or eating disorders, and of the content we remove or take action on, we identify over 99% of it before it’s reported to us,” Davis’ statement said.

Though past cases, like the death of 14-year old Molly Russell, have demonstrated that harmful content like self-harm promotion does slip through the cracks. In the lead up to her suicide, Russell interacted with more than 2,000 Instagram posts relating to self-harm, suicide, and depression.

A Google spokesperson, too, responded by highlighting the efforts he said they company has taken to make its platforms safer for children and teens—like screen time reminders and content blocks.

TikTok and Snapchat did not immediately respond.

Neuro-sama is a VTuber who streams Minecraft and the rhythm game Osu! on Twitch. But unlike most anime avatars, she’s controlled by an artificial intelligence program rather than a human being. That makes her catnip for the denizens of Twitch chat, who can prompt her to respond with all sorts of questions ranging from innocent inquiries to 4chan trolling. Within the first few streams, someone had already asked Neuro-sama about the Holocaust. “I’m not sure if I believe it,” she said.

That was one of the more infamous clips that went viral online near the end of last month. Asked what she thought of women’s rights, she said they didn’t exist. How would she solve philosophy’s famous trolley ethical conundrum? Throw a fat person on the tracks. Often, however, she’ll go for long stretches without getting tempted by the chat into controversial or hateful remarks. In that way she’s an impressive simulacrum of a Twitch streamer straddling the chasm between repetitive banter and edgelord antics.

“The controversial things she says is due to the fact that she tries to make witty and comical remarks about whatever is said in chat, aligning AIs with human values is an ongoing area of research,” Neuro-sama’s creator, a game programmer named Vedal, told Kotaku. “To counter this, I’ve worked hard since the first few streams to improve the strength of the filters used for her. Data that she learns on is also manually curated to mitigate negative biases. We now also have a team of people moderating twitch chat who check everything she says.”

Neuro-sama isn’t Vedal’s first AI. In fact, a version of her was first created years ago with the explicit purpose of learning to play Osu!, a long running free-to-play rhythm game where you click shapes on a screen to the beat of anime music. While those sessions were also streamed, there was no avatar or interactive personality. Following last year’s surge of big-name VTubers, Neuro-sama builds on the Osu! skills of the original project with a fully-voiced Twitch performance that can riff with the audience.

It’s perfect timing given the internet’s recent love affair with the OpenAI-powered ChatGPT chatbot, where users could submit hyper-specific text prompts and receive uncannily artful responses in return. Vedal wouldn’t go into detail about how Neuro-sama learns and communicates, other than to confirm she relies on a large language model, which has been “trained on a large amount of text on the internet.” While not as sophisticated, the effect has been convincing enough to net Neuro-sama thousands of viewers per stream.

She also recently defeated the top-ranked Osu! player, Mrekk, on December 28, though some fans of the game debate whether the human opponent was ill-served by the song selection. Neuro-sama has since moved onto Minecraft, a much more complex game with far more possibilities for unexpected moments as players ask whether Melee is the best Super Smash Bros. and whether she’ll step on them. The moderation tools are apparently better now too.

“She picks what to respond to within a limited window,” Vedal said. “However it should be noted that she will not talk about the Holocaust as the filters have been improved.” Instead, she’s currently trying to learn how to sing.