Tech Time by Tim #34

This week, the newsletter is extra chunky! Why??Because we’re going to talk about AGI! Not just that, but also about the Hollywood strikes and how they tie into longstanding productivity struggles between employees and employers. Moving on, we’ll also cover MSI’s disastrous hack, check back up on the Firefox Fediverse, and examine what Platforms that change owe to their original users. Last but not least, Microsoft x AMD, a story of Consolification and Consolidation.

At A Glance

A quick overview of this week’s content.

·??????The Week that Was: Work Wars – Generating Value And Measuring Productivity, AGI –Stairway To Heaven Or Highway To Hell?

·??????Cyber Safety Blanket: What The Hack MSI, The Firefox Fediverse

·??????Rules of Engagement: The OG, Microsoft x AMD

The Week that Was:

A look back at the tech world of the past week.

Work Wars – Generating Value And Measuring Productivity:

A historic writer’s strike is gripping Hollywood at the time of writing. For the first time in 15 years, the Writer’s Guild of America (WGA) has successfully launched a prominent worker uprising against major studios such as Disney, Netflix, Amazon, and Apple. WGA members accuse Hollywood of attempting to relegate human writers to editorial gig-work whilst shifting the actual writing over to generative AI.

At the same time, more than 40 leading scientists have resigned from the editorial board of science journal Neuroimage after owner Elsevier refused to lower its publication charges. Academics looking to be published in the journal must pay over €3000 for the privilege (per paper). The accusations levelled at Elsevier by the resigning academics are that the publisher’s business practices are unethical. Elsevier is estimated to have profit margins of around 40% and to control around 18% of all global academic publications. According to the accusations, Elsevier is parasitic in the sense that academics write the content, whilst the public funds them through government grants, for example. The way these two are connected is that it concerns the trump card of what I’ll refer to as ‘value generators.’ That trump card is the product, and by extension, productivity.

There is a paradox inherent to productivity. This paradox is grounded in the fact that no one truly knows how to define it. Thus, the more a company tries to quantify and measure productivity, the more time and effort employees have to expend to ‘look busy’ when they are not actively working. To still hit their targets, employees might send out so-called ‘productivity signals.’ Productivity signalling eats into work time. Thus, whilst productivity can seem high, efficiency might actually decrease and targets might be missed, In turn, managers might come to believe that employees with lower productivity scores must be the problem. Putting AI at the core of the user experience for productivity software might be a solution, but if it treats the symptoms or the cause of productivity issues remains to be seen. LINK

AGI – Stairway To Heaven Or Highway To Hell?

Here’s an interesting question to ponder. If you take the most serious threats posed by AI seriously, should you write about anything else? Yes! You should. In fact, that’s exactly what I’ll be doing in the rest of this newsletter. But first, let’s talk about Artificial General Intelligence (AGI). The ‘generally’ agreed upon milestone for AGI is that such an entity has human level intellect and abstract reasoning capabilities. As a result, it could theoretically be ascribed personhood, and therefore be held personally accountable for what it does. This is at once one of the most interesting and frustrating things about the AI ethics tug of war between evangelists and doomsayers.

I do not agree with the notion that AI’s emergent abilities are ‘just a mirage,’ but we do need to have more openness and honesty about what will power them. What I mean is that AI is, for the most part, human powered. It is a tool, and tools do not use themselves. Models do not train themselves, refine themselves, or maintain themselves. Humans do. Whether AI uplifts us, destroys us, or becomes a normal and ‘mundane’ part of our everyday lives…it will be because of human intent.

And so I choose to take both the apocalyptic and utopian possibilities seriously. Part of that is really thinking through the legal ramifications of SkyNet killing (almost) all of us. In that regard, AGI, too often gets used as a shield. If something goes disastrously wrong, it is too easy for (aspiring) developers of an AGI to absolve themselves of blame. Would we trial SkyNet for genocide once Sarah Connor and her plucky resistance have beaten it? No, the AI isn’t going to carry that weight dear developers, you are. In fact, this is in line with the historical roots of the word ‘robot.’ Karel ?apek's 1921 play R.U.R. (Rossum's Universal Robots) was the first instance of that word.

The robots in that only end up overthrowing humanity because humanity teaches them violence. In the end, two robots, mirroring Adam and Eve, inherit the earth and end the play on a positive note. Supposedly, it is derived from the word ‘Robota’ which means ‘drudgery’ in Czechian, and ‘work’ in the Slovakian. The beauty and integrity of the human spirit survived the supposed apocalypse, much like it did in 1927. Following the release of ‘The Jazz Singer,’ the so-called ‘Music Defence League’ campaigned to stop ‘evil machines’ from irreparably destroying art. Even though the evil machines won that round, yet art is still around today. It turns out that evil, quite like beauty, is in the eyes (and ears) of the human beholder. LINK

Cyber Safety Blanket:

Covering important cyber security developments so you sleep a little more soundly at night.

What The Hack MSI? – The Tenuous State Of Corporate Cybersecurity:

It’s been a pretty brutal week in the world of cybersecurity. Back in April, reports first began to surface of PC hardware manufacturer MSI being compromised by a hacking group called ‘Money Message.’ Their ultimatum for a $4 million ransom expired, and Money Message have now begun publishing the stolen data. That data includes source code, as well as the image signing private keys for 57 MSI products and Intel BootGuard private keys for 116 MSI products.

Intel BootGuard is a security system that prevents malicious firmware from successfully installing itself on PC’s. Although the amount of threat actors with the technical skills to do so is presumed to be relatively limited, they could in theory create malicious firmware updates which can pass the BootGuard safety checks on at risk MSI devices. At the same time, Western Digital, known for its memory storage devices, has also conformed a major breach. Customer data has been confirmed to be affected.

It's not just companies operating above board that had it rough though. Another large-scale international law enforcement operation took down ‘Monopoly Market.’ This dark web marketplace, active since around 2019, was primarily used for drug trade. I wonder how they ended up locating and arresting 300 people though, and the extent to which information obtained from organisations such as VPN providers might have played a role. From the way the Mullvad VPN police search went, getting user like that appears to be a lot more difficult than it might seem. LINK

The Firefox Fediverse – The Implications Of Mozilla’s Fakespot Purchase:

Mozilla is continuing its ‘fediverse,’ efforts. for those that have forgotten, the fediverse is a network of interconnected services. It has been around for quite some time but rose to prominence thanks to potential twitter alternative Mastodon. This past week it also acquired ‘fakespot,’ a plugin for the Firefox browser made and maintained by Mozilla. Fakespot detects fake reviews and other fraudulent content on popular marketplaces such as Amazon and eBay. Like web clipper service Pocket, Fakespot will likely be integrated into Firefox as core functionality in the near future.

The timing of this acquisition stands out for two reasons. Back in October of last year, the US Federal Trade Commission announced the start of an initiative to more comprehensively explore rulemaking intended to address fake reviews and other deceptive endorsements online. This past week the New York Times released a survey asking readers to share their experiences with such fraudulent content. Coincidence? Perhaps, but it is worth keeping an eye on.

It is worth keeping an eye on because successfully labelling fraudulent or otherwise problematic content is one of the more complex aspects of online content moderation. If Mozilla is able to get this right, it stands to gain a massive competitive advantage against rivals if it aims to keep pushing further into social media. It would certainly be an opportune moment, considering TikTok’s ongoing user espionage scandal appears to have gotten worse. LINK

Rules of Engagement:

Ethics and legal matters regarding tech engagement.

The OG – What Do Platforms That Change Owe Their Early Adopters?

The term “OG” stands for “original gangster” and refers to the earliest adopters of platforms or services. In many cases each username can only be taken once. Thus, having super short or obvious usernames is only possible if you choose them before anyone else that might want them. Disputes over usernames can get extreme. The reason people go to such lengths is because such handles are a mark of prestige and seniority.

The most common way to get around such issues is to add a numeric code after a name so that duplicates can coexist. You might be Example#1234 whereas I might be Example#4321. That’s how Discord has handled things… until now. Cynics point to the possibility that Discord may want to monetise any future verification features, the value proposition of which is greatly reliant on scarcity. Verification is a particularly thorny issue lately though. On Twitter, for example there are those who want to take a stand against Elon Musk’s approach. Others, such as sex workers, depend on it for their safety and livelihood.

Discord itself argues that as many as 50% of friend invites go to the wrong recipient because users don’t know their so called numeric ‘discriminator,’ or they make an uppercase vs lowercase mistake. But there’s more at stake than practicality. What tends to define the early days of social media platforms is a sense of shared identity and camaraderie.

The vibrant protocultures of ultimately successful platforms develop into unique ecosystems with their own cultures and subcultures. These ‘network effects’ are how ‘social’ spaces can become large enough ‘networks’ to sustain platforms. However, there’s an inherent conflict of interest between original users and the companies running the service. The bigger the service, the more money and market share the platform holder can earn. The bigger the service, the more addictive and toxic the battle for status and influence becomes for users. LINK

Microsoft x AMD – Consolification and Consolidation:

I was wrong last week! Kind of. Microsoft is indeed going after AMD… but as suitor, not an assassin. It’s a partnership for now, but I do wonder about that in the long run. Microsoft is definitely the bigger fish here. Having said that, the partnership is mutually beneficial for now. At face value, of course it’s about making an AI chip. But that’s not the interesting part about this particular duo teaming up.

AMD can’t beat NVIDIA on its own, and Intel, despite having plenty of fires to put out at present, still isn’t a pushover by any means. Whilst absolutely a force to be reckoned with on the software front, Microsoft’s hardware successes have been… well… yeah… so bad it has finally, openly admitted its console war defeat by Sony. So, let’s talk a little more about ‘Consolification.’ Did you know that AMD hardware powers the PS5? I also mentioned last week how many steam deck alternatives were powered by AMD. And this pseudo-standardisation is a part of the Consolification phenomenon.

PC optimisation is hell in part because there’s countless different unique configurations across millions and millions of users. My devil’s advocate duties thus concluded, I have to say that PC performance of cross-platform titles has been particularly shocking of late. So it might seem weird that new figures for Activision’s games appear to show them making more money on PC than on console, now of all times. The gap is even widening. Could it have something to do, perhaps with the rise of social gaming? Though initially more of a mobile thing, Microsoft’s game pass is far from the only service to cross the mobile/PC barrier, whilst console ecosystems remain more insular and walled off. LINK

A Nice Cup of Serendipity:

Cool bits and bobs from around the web.

Food Art LINK

Deep See Baby LINK

Lightsabers LINK

Regrowing Limbs LINK

Google I/O LINK

Fallout History LINK

Unexpected Pollinator LINK

Schr?dinger’s Chonker LINK

Retirement LINK

Disunity LINK

Nomophobia LINK

Blue Blues LINK

CS:GO Info LINK

Edging vs. Chrome LINK

Fortnite Olympics LINK

Look Ma, No Screens! LINK

The Deep End:

A weekly batch of longform content recommendations.

An AI Artist Explains His Workflow:

An example of generative AI used as one of the tools in an artist’s toolkit. LINK

Hive Mind:

A Beekeeper explores her relationship with her hive and how it evolved over the course of the pandemic up until now. LINK

A History of NASA’s Super Computers:

A thorough look at the how early computing evolved to make historic events like the lunar landing possible. LINK

Microsoft’s Digital Defence Report For 2022:

Though obviously intended to make Microsoft themselves look good, this report is nonetheless an extremely in-depth look at the overall cybersecurity threat landscape. LINK

We’re not remaking horror games, We’re chasing nightmares:

A critical reflection on the present state of the horror game genre. ?How its biggest recent hits have been old games instead of new ones. LINK

One More Thing…

You know the song “push it to the limit” that plays during the climactic final showdown in Scarface? Yeah, that was me skirting up against the word limit that LinkedIn will allow for its newsletters. At the risk of a copyright claim, here’s my spin on it:

Push it to the limit. Type along the razor’s edge, don’t back down, just keep going or you’ll be finished. Hit space bar and double the stakes, tabs wide open like a bat out hell, you crash the site, crash the siiiiiiite! Going for the back of beyond, nothing’s gonna stop you there’s nothing that strong, so close now you’re nearly at the brink so push it! Oh yeah!

And push it I did. I’m a little too passionate about tech ethics, so making something as complex as AGI morality and accountability fit into a single one of my items was pretty taxing. See, the thing with AI is that it’s like a wrap. You can put practically anything in it. Mix ingredients, remix dishes you’ve already had a different way previously. The possibilities are endless. But you know… That wrap isn’t going to wrap itself. Someone, somewhere, also had to prepare those ingredients you’re using as well. And if you try to stuff too much into your wrap…you’re going to need to work a little harder to make it all fit.

I actually typed this ‘one more thing’ on Tuesday night at around 23:00 PM. Why was I working so late? Because you see dear readers, I did not need sleep, I needed answers. To what? Well, I often get asked about the other things I do, aside from the newsletter. Honestly, that question is a real struggle to answer. How the hell am I going to tell people with a straight face that I lie awake at night thinking about whether ChatGPT dreams of electric enchiladas?