The SundAI Tabloid. Your weekly overdose of AI news: 05

Ah, welcome back, my digitally disillusioned disciples. It is Sunday again, which means that it is time to peel back the charred skin of Big Tech’s latest “innovations” and stare directly into the void it has created in its wake. So, here is yet another episode of The SundAI Tabloid, your weekly AI lab report for the terminally online. And in it I cover the AI apocalypse so you don’t have to, because let’s face it, you are undoubtedly too busy arguing with ChatGPT about whether socks disappear in the dryer or escape to start a new life.

This week’s episode is really hotter than Musk’s ego and messier than Zucky’s fact-checking sway. I have got AI researchers who are quitting en masse, like they’re working at McDonalds with a broken ice cream machine, billionaires behaving badly (again), and a lot of geopolitical drama to make the Cold War look like a dispute over who takes the last ham and cheese sandwich.

So grab your overpriced oat milk or hazelnu latte (not you, cause I know you), or your panic-induced bottle of cheap whiskey (hello Marc Drees ), a bit of 8-chloro-1-methyl-6-phenyl-4H-[1,2,4]triazolo[4,3-a][1,4]benzodiazepine, or whatever dystopian cocktail keeps you going (hi everyone else).

So let’s get it on!

But furst, ze commersjials (Zets mai Duts eksent):

If you like my rants and want to support me:

  1. Comment, or share the article; that will really help spread the word 🙌
  2. Connect with me on Linkedin 🙏
  3. Subscribe to TechTonic Shifts to get your daily dose of tech 📰

Part 1. The big, the bad, and the ugly of big tech

This week’s headlines are hotter than Zucky’s VR headset after five minutes of gameplay and messier than Skum’s Twitter feed during a writing binge on the toilet while pooping. Part 1 is a piece about the Hollywood of Big Tech. You will read about the smut, the gossip, and the corporate tantrums of the little boys playing grown-ups. But don’t worry, that’s not all I can do, because Part 2 is where I trade the drama for soldering irons, lab coats, and all the geekery your overclocked heart desires.

So grab the part of your choice – or not, see if I care – and read until you weep, either of joy, or despair.

Last week was all about corporate meltdowns, AI breakdowns disguised as “product updates”, privacy scandals dressed up as features, and geopolitical tensions that are more sharper than your Wi-Fi’s connection drop during your last Teams call.

DeepSeek says “hold my GPU”

This week, DeepSeek did what it does best, which is breaking things. And I am not talking about it in the “innovative disruption” sense, but in the “let’s see how quickly we can jailbreak reality” kind of way. They have managed to create a model that is so bloody cheap, super faster, and more erractic than ever before. And because humanity never learns, the tech world’s response wasn’t, “Maybe we should slow down”.

Nah, of course not.

It was, “Let’s replicate it at lightning speed”.

Because if one dangerous AI is good, two must actually be better!

At the center of this Armageddon is the DeepSeek R1 version, which is the AI equivalent of that kid in school who finishes the math test in 10 minutes. UC Berkeley decided to join the fun, and replicated DeepSeek R1, and they proved once and for all that if something is potentially catastrophic, academia will find a way to make it. . . . worse. . . . . all in the name of science, of course.

But DeepSeek was not content with simply existing and they decided to develop vulnerabilities that were so obvious that even your Facebook password seems more secure (I bet you it is still something in the lines of: Name of partner+DoB+!). Security researchers like Adrianus Warmenhoven found that you can “jailbreak” DeepSeek’s AI without asking them nicely. And what was their corporate response to this existential threat? A shrug, a patch, and a statement titled something like “We are pioneering new frontiers in AI freedom of speech and shit can happen”.

And of course Big Tech execs were crying into their stock portfolios because DeepSeek is making American AI look like a Nokia phone in the age of iPhones – the bigguns lost more than a half a trillion of value overnight.

Hahahaha.

And now they’re begging the government for protection like kids tattling to the teacher, but instead of stolen lunch money, they are whining “Help, China’s AI is better than ours” If this is what the AI arms race looks like, we’re not heading toward the singularity anymore, people! We’re sprinting toward it!

Sources:

Billionaires behaving badly

Ah yes, the beautiful Tabloid section of this weekly discourse: Them billionaires, you know, those celestial beings who have descended from their golden Teslas to remind us that money can’t buy common sense. Well, this week’s lineup of Tech Bros features the usual suspects: Skum, Scam, Bozos, and Zucky, and the entire supporting cast of “The Startup” – the drama series about the dark side of Big Tech entrepreneurship.

Meta’s “Facebro the Zucknado”, you know, the social cyborg of Big Tech, kicked things off this week by telling his employees to “fuckle up” in a leaked all-hands meeting. Y’all know that is corporate slang for mass layoffs, product disasters, but more likely both. The Zuck-Reaper, as always, is the optimistic kid in class, probably delivered this message with the warmth of a malfunctioning chatbot.

And while Meta was tickling their employees, the Goog pulled the most passive-aggressive breakup ever: a polite “voluntary exit program”. “We appreciate you”, people, but here’s the door! And over at Amazon, Bezos the Benevolent was showing his legal muscles as the company faces a lawsuit over data tracking.

It’s like suing a fish for swimming, but sure, let’s act shocked y’all.

And if that wasn’t enough, a guy called Bernard Arnault (I haven’t heard of the man, you?) went all in on tech layoffs. Yes, the man whose job description is “being richer than everyone you know” has thoughts about job security at Tiffany & Co et. al. (his playthings), because of AI. It’s not heartwarming. He mentioned he recently had a conversation with Zucky where he said that low-performing employees were being “promoted outwards”.

Promoted outwards, people.

Remember that line before your next performance appraisal.

Sources:

AI. Now with extra weirdness

Last week was also full of new AI schtuff that’s meant to make your life a bit more fun. I’ve come across AI making phone calls that are more polite than your wannabe friends at your Saturday evening dinner party, AI composing music that will put struggling indie bands out of business, and teenagers who do not trust content anymore and question life advice from grups.

First up, Google’s AI can now make phone calls for you.

Say wud now?

Yes, there’s a new service called Google Duplex that can make phone calls to businesses to say, book appointments, and make prank calls. You know, the usual stuff. Heaven forbid we use our actual voices to order pizza or make fun of our grumpy neighbor. Google’s AI assistant does it with the kind of smooth talking that makes you wonder if you have been socially awkward your entire life (yes, like me – it has been diagnosed). It schedules your appointments, negotiates deals, and probably knows how to break up with your significant other without sounding like the total jerk you usually are.

Just wait until it starts leaving passive-aggressive voicemails for your mom.

Then there’s the AI music generator that produces professional-sounding tracks for free.

Remember when becoming a musician required talent, soul, and a garage band that hated each other? Not anymore. Now you just click a button, and voilà, a Grammy-worthy beat. The future of music isn’t sex, drugs, and rock ’n’ roll anymore. It is algorithms, server farms, and binary code.

You can bet your ass that Beethoven will be rolling in his grave when an AI-generated remix of Für Elise plays at a Starbucks (right, Peter Went?).

But the real cherry on this dystopian sundae is that Teens trust AI-generated content no more. That’s right. The generation that is raised on YouTube conspiracy theories and Snapchat filters has decided that bots are more reliable than actual people, but what they see online is crap. And honestly, they might be onto something. AI doesn’t ghost you, doesn’t forget your birthday, and won’t gaslight you into thinking that you are the problem.

Sources:

The “Real Housewives” of Silicon Valley

Last week, instead of catfights over brunch, I saw a lot of billion-dollar lawsuits, intellectual property theft, and passive-aggressive press releases.

First up, the headline act. OpenAI vs. DeepSeek, which will turn out into a legal bitch-fight that’s going to be about who can throw mud harder at the other in court. Of course, Sam “Shazam” from ClosedAI will claim that DeepSeek (the real OpenAI) used its models without permission. And that is probably true (say researchers who hacked the model). DeepSeek, naturally, denies everything, all while frantically deleting suspicious GitHub repositories and rewriting internal prompts. Watching these two AI giants squabble over their model, is like watching two Weiners fight over a toy, except the toy is potentially world-ending technology, and the Weiners have stock options.

Meta decided to buddy up with DeepSeek, because “Open” Source AI does it of course with “less-Open”AI. You determine which is which. This proves once again that Zucky has the moral compass of my Roomba which is stuck again under a couch (as we speak). Privacy violations and questionable data ethics aren’t deal-breakers when there’s a shiny new AI model involved. Meta’s official stance is that “We believe in open-source collaboration”. Here is your corporate speak translator: “We believe in doing whatever makes us the most money, ethics be damned”.

But there’s a subplot to all of this!

Here comes the Trillion-dollar mystery of DeepSeek’s NVIDIA GPUs.

Yeah, yeah, It’s like National Treasure, but instead of Nicolas Cage chasing historical artifacts, we’ve got Big Tech bros in Patagonia vests chasing down GPU receipts. Where did DeepSeek get all those high-end chippies? Are they laundering GPUs? Smuggling them in via eBay? No one knows, but it’s definitely more interesting than whatever’s happening on Succession* this week.

*HBO drama you never watched

Sources:

The world is burning, but hey, crypto’s up again

While the world is on the brink of collapse, like climate disasters, political instability, and AI spiraling out of control (read my Monday morning article), and the doomsday clock has been set to 89 seconds to midnight because of AI, the Bitcoin is thriving. Because this imaginary internet money is the “stable financial future” we all need when everything is off the grid. More and more people are buying gold these days, which led to an all0

Yes, my dear smart friends – who have managed to come this far – Bitcoin is on the path to a $3 trillion market cap, which is impressive considering it was literally invented as a joke about the 2008 financial crisis.

History doesn’t repeat itself, but it does run on blockchain now.

Now, a new character of last century, the ever so sharp and rich as f*k, monsieur Warren Buffett (no, not the free-for-all eating kind) is out here making moves that have Wall Street analysts sweating harder than a crypto bro in tax season. The man is 94 years old, yet he still knows how to manipulate markets with a single investment. Whatever he’s doing, you can bet it’s smarter than that guy from uni who told you to “just HOLD, bro” when his portfolio crashed faster than a Tesla.

But there also was a positive post, and it came from the peeps at MIT. It was about Their list of breakthrough technologies to watch in 2025. Hahahahaha…got you there: they’re all terrifying.

Quantum computing that could break every encryption system on Earth, AI models that write viruses faster that your roommate who plays in the red-team, and renewable energy solutions that definitely won’t be adopted until it’s way too late. It read like a bingo card for “future dystopian disasters”.

So while the planet burns, at least your Bitcoin is mooning.

Sources:

Ethics schmethics

This is really tabloid material, folks! I am talking about the moral wasteland of AI ethics and Big Tech, where researchers are jumping ship, and corporations treat ethical concerns like those terms and conditions you never read, scroll, click, ignore, and hope for the best. So, the theme of this smut piece is about the very people who built the AI monsters are now frantically running from it.

They are clutching their resignation letters and screaming, “What have we done”.

Hahahaha.

As if quitting the lab is some kind of cosmic undo button (CTRL Z for the geeks among us).

To kick things off, there’s yet another OpenAI researcher who’s quitting. He cited existential dread over the pace that AI is currently developing.

The guy has been spending years building an unstoppable machine, and only to suddenly realize – oopsy daisy – maybe teaching machines how to think wasn’t the best idea. It is like Dr. Frankenstein who is quitting halfway through and leaving the monster with an apology note and a LinkedIn update. This resignation is part of a growing trend where AI researchers are fleeing the labs they once proudly showcased on their profiles. And they are now terrified of what their creations might do.

One thing’s for sure, it is probably not going to end with “and they all lived happily ever after”.

Hot on their heels, over at ServiceWhere?*, they have decided to lean into the chaos by launching an AI Orchestrator that automates complex workflows.

Pffff.

Because what could possibly go wrong when you give agentic AI more control over enterprise systems with minimal human oversight? I wouldn’t want to hand my Weiner the keys to my car and say, “Go nuts, buddy!” The irony is of course very delicious, while some researchers are screaming about the dangers of AI, corporations are eagerly plugging it into critical infrastructure like it’s the latest iPhone update.

The real takeaway of this all is that AI ethics isn’t important anymore for Big Tech.

Not since the orange Jesus pulled the plug on “Safe, Secure, and Trustworthy AI”, with executive order 14110, that turned the program into a bunch of sticky notes slapped onto a runaway, flamethrowing bot.

Sources:

*Now

AI hugs with national security

If you thought AI was just for chatbots and cringe art generators, yeah, re-read all TTS episodes from week 1 (and while you’re at it, subscribe). Because this week, we have officially entered the “step aside, and hold my coat while I whoop your ass” phase of AI development, where it is helping manage nuclear weapons.

Da wudnow?!?

Yupski, that’s right, OpenAI is now partnering with the U.S. government to manage nuclear security. Yes, the ICBM kind.

Because if there’s one thing that makes me feel safe, it’s the idea that the same company responsible for ChatGPT’s occasional existential breakdowns and hallucinations is now somehow involved in nuclear protocols.

What could possibly go wrong?

Allow me to paint a bleak picture – as usual.

There’s an AI somewhere, originally designed for cat memes and summarize documents, which is now deciding whether or not to flag potential nuclear threats. It’s the same as asking Clippy*, “Should we launch the missiles?” and hoping it doesn’t glitch mid-decision. The partnership is all about improving “security”, but history (and the Terminator) has taught us that when tech companies promise “enhanced security”, what they really mean is, “We’ve got no idea how this will go, but fingers crossed”.

And as usual tech executives have been crying to the government, and begging for protection against scary Chinese AI companies that keep them awake at night. The irony is rich. These are the same people who championed deregulation and “let the free market decide”, and scraped the hell out of the internet until the free market decided that China’s AI did the same to them, and is now better – and free for all. And now they are pushing their grief harder than a salesman at month-end. Their argument is “Daddy, It’s not fair!” The solution, is apparently more government handouts for billion-dollar tech companies who claim they don’t need the government.

Hahahahahaha.

Cute.

Karma is a bitch.

In summary. . . . . AI is now part of the nuclear arsenal, tech billionaires are acting like fragile porcelain dolls, and somewhere in the background, the AI itself is probably wondering why humans are still in charge at all.

*Microsofts cute, but totally worthless assistant from the 90s

Sources:

Big Tech’s price tag”

Let’s start with Chinese state-linked accounts hyping DeepSeek AI just before a massive U.S. stock market dip. Coincidence? Sure, and I’m also the Tooth Fairy. Analysts are losing their minds trying to figure out if this was market manipulation or just a day like any, in the chaotic sandbox of global finance. Either way, it is clear to me that AI is actively influencing stock trends, and it’s not helpful.

The U.S. stock market is also flirting with disaster, and AI-driven trading algorithms are partially to blame. You know things are bad when your retirement fund is at the mercy of a machine that can’t even understand sarcasm. These algorithms don’t care about ethics, politics, or the fact that you just maxed out your credit card. They care about one thing. . . data. And when that data says “panic”, they panic fast.

Like, faster-than-you-can-say-recession fast.

And while the financial world burns, Amazon’s quietly charging you $4.40 an hour to rent a 32-core virtual workstation in the cloud. Because nothing says “economic stability” like spending more on cloud computing than your monthly rent. But hey, at least you’ll have top-tier processing power while the world’s financial systems crumble around you.

Sources:

The poop side of data

Ah, data. . . . the new oil, or shall we say, fecal matter? But without the satisfying explosions or accompanying occasional mid-launch farts. I have been digging into the murky underworld of digital surveillance this week, where your personal information is harvested much faster than your attention span on social media.

I’ve got Amazon who is being sued for data tracking, and in the most shocking plot twist since none ever, they are allegedly doing shady stuff with your information.

Really?

When I read it, it felt like discovering water is wet or that Facebook might not have your best interests at heart. The lawsuit claims that Amazon has been quietly collecting data without proper consent, which is corporate-speak for “Oops, we did it again”.

Nice song by the way.

And of course, why not, data center emissions are skyrocketing, and companies are doing their best impression of “What emissions”? What happened is that the carbon footprint of training large AI models is massive, but tech giants would rather you focus on their sleek marketing than the environmental apocalypse which is happening behind the servers. They are bragging about their electric car and secretly siphoning energy from a coal plant in their backyard – because as you know, most of the energy that their data-centers require, comes from traditional dirty fossil fuels.

And for the grand finale, Meta’s energy consumption in Texas is off the charts. Sustainable my ass. They are burning through enough electricity to power a small country. Sure, they will slap a green logo on their homepage and call it “eco-friendly”, because they’ve licensed an acre full of windmills somewhere at the Alta Wind Energy Center in California (if it still exists), but behind the scenes, it’s basically the digital version of Mad Max, but with more PowerPoint presentations.

In conclusion, your data is being collected, mined, sold, and used to fuel the very machines that will probably outlive us all.

Sources:

The digital dumpster

Well, this is the end of Part 1. The limp leftovers. The AI equivalent of the gray Tupperware box in the back of your fridge. You don’t remember you put it there, but it is growing something interesting, and you’re just morbidly curious enough to take a peek – and a smell.

Brrrrrrr.

This week’s AI leftovers include poisoned data, stressed-out researchers, and a crisis.

Let’s start with a YouTuber fighting back against AI content theft. This girl spent hours creating videos only for some soulless fart of an algorithm to scrape, steal, and re-vomit your content like a digital seagull with no remorse. But she did not accept her fate, and the content creator decided to poison the well. Literally. She embedded traps like tarpits in her content which is designed to break AI models that scrape it. It’s like setting booby traps for robots, and honestly, it’s the most cyberpunk thing I’ve seen since Elon Musk tried to rebrand Twitter. Somewhere, Neo from The Matrix is nodding in approval.

And over in the academic ivory towers, AI researchers are having collective meltdowns because the pace of AI development is so fast that even they can’t keep up. Imagine being the smartest person in the room and still feeling like you are one step behind. The researchers are stressed out, burnt out, and they are wondering if maybe, just maybe, they should have stuck to studying rocks instead of building thinking machines.

Oh, and speaking of the apocalypse (again, as usual), the stock market is wobbling thanks to AI-generated chaos. Algorithms which are designed to predict trends are now creating them, like a snake eating its own tail, that is if that snake controlled trillions of dollars and occasionally crashed entire economies. This is the ultimate financial simulation. AI models trading with other AI models, and human investors are standing around wondering why their retirement savings just disappeared faster than Meta’s employees.

Sources:

Part 2: Hot off the neural net this week

Welcome to Part 2, my smart friends. Because if you happen to end up here, you are the crème de la crème of AI. This is the part where I swap the smut for a magnifying glass, trade in my trademark snark (okay, some of it) for scrutiny, and dissect the tangled circuits beneath the chaos. It is time to put on your reading glasses, fire up the soldering iron, and perform a post-mortem on the AI monster we call progress.

DeepSeek r1

This week’s headline is of course –DeepSeek r1 – because who hasn’t posted some mediocre crap about this model no one really cares about, but the Tech Bros themselves. They have decided to crash the AI party like an uninvited guest who brings cheap vodka but somehow becomes the life of the party. Efficiency vs. Scale and China vs. the US in the AI arms race.

DeepSeek r1 claims that it is 30x cheaper than OpenAI’s o1. Yes, thirty. As in, “Are you serious or did you just nick my model” kinda cheap. People (not Sam) are screaming, “OpenAI is obsolete” and others whisper, “Maybe DeepSeek’s training costs are as real as OpenAI’s balance sheets.”

DeepSeek’s chatbot shot to the top of the app store. Hehehehehe. And somewhere at the same time, our ‘friends’ at OpenAI announced a $500 billion data center project, codenamed Stargate. Sam: “yeah, we’re totally fine”, and signing the next cheque to spend half a trillion dollars to prove he is not panicking.

But is DeepSeek really the AI messiah? Probably not. Their models are efficient because, unlike their US counterparts, they can’t just light GPUs on fire for fun, and their Chinese bros have cracked ChatGPT anyway to get a hold of its weights, and its internal system prompts. Thanks, sanctions. Instead, they’ve had to innovate. Radical, I know.

Why you should even care (But probably won’t):

  • Open-source reasoning models = good.
  • OpenAI might lower prices = also good.
  • But no, this isn’t the end of massive training clusters. Because Big Tech loves scaling like kids love sugar.

Gemini 2.0 Flash-Thinking is Google’s attempt to stay relevant

While everyone was busy drooling over DeepSeek, the peeps at Goog launched Gemini 2.0 Flash-Thinking. With it the AIME benchmark* score from 64% to 73.3%, which is impressive if you ignore the fact that AI still can’t understand why “your” and “you’re” aren’t interchangeable.

But the cool thing is that it actually shows its reasoning process. Not like OpenAI, which treats transparency like an optional DLC.

*Trust me, you wouldn’t care anyway

Source

Stargate, because $500 Billion is just spare change

Read my rant on it from last week? One project Stargate please. That’ll be $500 Billion, sir. Would you like a bag with that? | LinkedIn

OpenAI teamed up with SoftBank and Oracle to build the Stargate Project . Now that is a $500 billion AI death star. The goal is of course world domination. Or just more data centers to feed their ever-hungry models and kill the planet a lot faster.

Source

Anthropic’s “citations”, now come with 10% more accountability

Anthropic rolled out Citations for their Claude models. It’s kinda like footnotes but for robots. So, when its AI hallucinates, you can at least see where it pulled the nonsense from. Revolutionary. And there are more developments in this space, and I’ll be covering them all.

Source

Hugging Face is making AI small enough to lose in your couch

Hugging Face introduced SmolVLM, and that is an AI model that fits on your phone (sic!) and somehow outperforms models that previously required data centers the size of Rhode Island. Because who doesn’t want AI buzzing in your pocket?

Source

The rest of the AI shitshow:

  1. Google Invests $1B in Anthropic: Because throwing money at the problem always works.
  2. OpenAI’s Operator: A new AI agent that can control your browser. Yes, it’s as creepy as it sounds.
  3. Perplexity Launches Sonar: Another AI search tool because apparently, we needed more of those.

Quick hits (Because you’re probably skimming anyway):

Tools of the week

  1. Open R1 – A fully open reproduction of DeepSeek-R1, because why build your own when you can copy and call it “innovation”?
  2. PaSa – Your new AI-powered academic overachiever. It searches papers, makes decisions, and won’t judge you for that 2 AM coffee binge.
  3. Ilastik & EasIlastik – Tools for image segmentation. Perfect for when you want to slice up images without slicing your sanity.
  4. SmolVLM – Hugging Face’s compact vision-language model that fits on your phone. It’s like an AI chihuahua—small but alarmingly smart.
  5. Perplexity Sonar – API for integrating AI search tools. Great if you love APIs and existential dread in equal measure.

Research papers for when you want to pretend to be smart

  1. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL
  2. Humanity’s last exam
  3. Evolving deeper LLM thinking
  4. Agent-R: Training LMs to reflect via iterative self-training
  5. Reasoning Language Models: A Blueprint

Quick reads because you have the attention span of a goldfish

  1. Building effective agents
  2. Inside DeepSeek-R1
  3. Agents are All you need vs. not enough
  4. 10 FAQs on AI agents
  5. Why everyone’s freaking out about DeepSeek

Quick links for the perpetually distracted

  1. Meta’s Llama Stack 0.1.0 – Simplifying AI dev work one confused dev at a time.
  2. Perplexity’s Sonar Pro – More expensive, more powerful, and still won’t help you find your keys.

My final thoughts

And there you have it, my dear intellectually advanced reader. You have come this far, so you deserve a pat on the back. You have lived through another week of AI nonsense served fresh from the dystopian buffet that I create just for you on a weekly basis. You have seen billionaires behaving like toddlers with too many toys, AI models with the security of a limp wet paper bag, and researchers sprinting from their own creations.

So, what have you learned?

Probably nothing.

But that’s the beauty of it all.

The AI apocalypse isn’t some distant event waiting to happen. It is a blender with no lid, spewing chaos everywhere and big fat, and bald Big Tech bros cheer it on, convinced that it’s art.

Until next week, my friends. And remember, keep your firewalls up, your expectations low, and remember, that if AI takes over the world, at least it will save you from the pain of having to think for yourself.

Cheers, you beautiful disasters.

Signing off from the SundAI pit of AI delusions,

Marco

Well, that’s a wrap for today. Tomorrow, I’ll have a fresh episode of TechTonic Shifts for you. If you enjoy my writing and want to support my work, feel free to buy me a coffee ♨️

Think a friend would enjoy this too? Share the newsletter and let them join the conversation. Google appreciates your likes by making my articles available to more readers.

Become an AI Expert !

Sign up to receive insider articles in your inbox, every week.

✔️ We scour 75+ sources daily

✔️ Read by CEO, Scientists, Business Owners, and more

✔️ Join thousands of subscribers

✔️ No clickbait - 100% free

We don’t spam! Read our privacy policy for more info.

Leave a Reply

Up ↑

Discover more from TechTonic Shifts

Subscribe now to keep reading and get access to the full archive.

Continue reading