BlueMonday1984 @ BlueMonday1984 @awful.systems

Posts

90

Comments

975

Joined

2 yr. ago

I still remember the outcry when Harambe was shot. Shit truly does feel like the point where everything began turning to shit.

Someone I know argued that Harambe's death led to Trump's election, and its been burned into my mind ever since:

To give some slack to HitchBOT, the design at least had some personality to it. That is a lot more than you can give the average tech-related thing these days, especially in an age of AI slop.

By my guess, the servers and datacentres powering the LLMs will end up as the AI bubble's Range Rover equivalent - they're obscenely expensive for AI corps to build and operate, and are practically impossible to finance without VC billions. Once the bubble bursts and the billions stop rolling in, I expect the servers to be sold off for parts and the datacentres to be abandoned.

New Baldur Bjarnason: The melancholy of history rhyming, comparing the AI bubble with the Icelandic banking bubble, and talking about the impending fallout of its burst.

Found a Pivot to AI candidate in the wild: Pentagon Document: U.S. Wants to “Suppress Dissenting Arguments” Using AI Propaganda

I also found a call for ethics training in engineering, and someone's horror story about ethics training alongside it.

I'd just like to say congrats on making it into NYT - it took 'em long enough to recognise you were worth listening to.

AI bros keep being literally unable to tell good writing from bad writing, so they tell you that obvious slop is just fine when it really isn’t. But editors can tell. Do not write like a slop machine.

Going on a tangent, I can see English/Creative Writing degrees getting a major boost in job market value thanks to that being exposed - on top of showing you don't need spicy autocomplete to write for you (which I predicted two weeks ago), getting such a degree also shows a basic ability to tell good writing from bad writing.

Before the bubble, employers could easily assume anyone they hire would be capable of telling good writing from bad writing by default. Now, they the possibility of a would-be hire being incapable of even that basic feat is something they have to contend with.

I was planning to mention Procreate as well, but felt like that'd be spamming the replies a bit.

On a wider note, I expect it'll be primarily art-related software/hardware companies that will have avoided AI participation - with how utterly artists have rejected the usage of AI, and resisted its intrusion into their spaces, the companies working with them likely view rejecting AI as an easy way of earning good PR with their users, and embracing it as a business liability at best, and a one-way trip past the trust thermocline at worst.

Gonna cheat a little bit and put one-woman consultancy firm/personal blog deadSimpleTech up as an example. The sole member is Iris Meredith, whose involvement begins and ends at publicly lambasting AI's continued shittiness.

Starting things off with a fun post I found on Bluesky:

the most productive way to do things is to do it deliberately and with good planning

Two things which coding is currently allergic to, as the rise of vibe coding has demonstrated

Public reminder that two thirds of American Jews support the Gaza genocide. ScottA is not an outlier, he's the norm.

Assembly: really gets you to understand that you are contending with a computer chip, and that anything interesting that you want to do requires abstraction.

This is only tangential to your point, but I did remember (now-defunct) game studio Zachtronics put out a few games heavily featuring assembly: TIS-100, which directly revolves around programming the titular computer in its own version of assembly, and SHENZHEN I/O, which centers around building embedded systems and programming the microcontrollers contained within.

The company's catalogue is completely free for schools under the Zachademics program, so you could use them to show how assembly programming's like if you were running a school.

CIO even ends with talking up the Luddites — and how they smashed all those machines in rational self-defence.

I genuinely thought this wasn't true at first and went to check. Its completely true, a fucking business magazine's giving the Luddites their due:

Regardless of the fallout, fractional CMO Lars Nyman sees AI sabotage efforts as nothing new.

“This is luddite history revisited. In 1811, the Luddites smashed textile machines to keep their jobs. Today, it’s Slack sabotage and whispered prompt jailbreaking, etc. Human nature hasn’t changed, but the tools have,” Nyman says. “If your company tells people they’re your greatest asset and then replaces them with an LLM, well, don’t be shocked when they pull the plug or feed the model garbage data. If the AI transformation rollout comes with a whiff of callous ‘adapt or die’ arrogance from the C-suite, there will be rebellion.”

It may be in the context of warning capital not to anger labour too much, lest they inspire resistance, but its still wild to see.

"Would you like code with that?"

That was where the "creating a mirage of it" link was earlier. Removing now.

what to do with this information

If you know any sci-fi/fantasy mags, you should probably tell them about it to help them identify and reject slop more easily.

A pull request is when someone submits new code to a software project. On 21 August, NX added some configuration to look at the titles of pull requests and check they were correctly formatted.

I find it immensely hilarious that this security hole was blown open on my 25th birthday. Its almost poetic.

By my guess, its gonna take about a decade to fully clean up the mountains of slop code that this AI bubble's gonna leave. It'll certainly be lucrative (and soul-deadening, as you note), but as someone else has noted before, the riches are exclusively going to experienced devs and senior programmers - for anyone trying to break into the industry, they're probably gonna have to find work somewhere else.

New Pivot to AI candidate just dropped: Taco Bell rethinks AI drive-through after man orders 18,000 waters

Found a couple articles about blunting AI's impact on education (got them off of Audrey Watters' blog, for the record).

The first is a New York Times guest essay by NYU vice provost Clay Shirky, which recommends "moving away from take-home assignments and essays and toward [...] assessments that call on students to demonstrate knowledge in real time."

The second is an article by Kate Manne calling for professors to prevent cheating via AI, which details her efforts in doing so:

Instead of take-home essays to write in their own time, I’ll have students complete in-class assignments that will be hand-written. I won’t allow electronic devices in my class, except for students who tell me they need them as a caregiver or first responder or due to a disability. Students who do need to use a laptop will have to complete the assignment using google docs, so I can see their revision history.

Manne does note the problems with this (outing disabled students, class time spent writing, and difficulties in editing, rewriting, and make-up work), but still believes "it is better, on balance, to take this approach rather than risk a significant proportion of students using AI to write their essays."