#mole — Public Fediverse posts

In giro per l'Italia, Mole Antonelliana - Museo del Cinema, Torino.
--
#travel #torino #momenti #traveldiaries #viaggi #italia #visioni #Foto #Fotografia #museodelcinema #mole #photography #photo #pic #travelphotography #veroexplore #behindthelens #shot_italia #igers #alessandrogaziano #pictasmonochrome #biancopiùnero #blackandwhite #snap_bnw

The need to use a #MOLE to solve an OS problem is a *symptom*. Of an architecture or UX problem that needs solving, so people don't need to resort to such desperate measures to operate their computers.

The risk of normalising the use of MOLEs to hack around symptoms like this, is that it means the deeper problems stops being seen as a problem, and doesn't get fixed.

(2/2)

A classic example is KiwiBank, a publicly-owned banking corporation in Aotearoa. As NZ citizens, we're co-owners, not nonpaying customers. But we still can't use their online services without becoming the product, by using their proprietary apps.

The one on the website pulls scripts from a number of third-party domains, owned by #MOLE Trainers, and doesn't work when those domains are disabled (eg by NoScript). I don't use the mobile app but I'm guessing it's the same or worse.

(3/3)

#KiwiBank

Even more shocking, although not really a new problem specific to #MOLE Training;

"This article examines digital forced labour through a human rights lens, analysing its characteristics, legal implications, and regulatory challenges, with particular attention to vulnerable populations, including children."

#RohitaashvSinha, 2026

https://ksandk.com/labour-employment/digital-forced-labour-in-ai-legal/

Some of @pluralistic's fiction references labour organising against this .

#ChildLabour #DigitalForcedLabour #ForcedLabour #ModernSlavery

Mole – Deep research agent for your terminal

https://github.com/lajosdeme/mole

Comments: https://news.ycombinator.com/item?id=49303046

#HackerNews #Mole #Deep #Research #Agent #Terminal #GitHub #Productivity

I'd like to see a #MOLE add the Public Domain Mark to any software it vomits up in response to a prompt, with no human editing the code. Reflecting the fact that no one has a copyright monopoly over completely auto-generated software. As well as informing anyone trying it that it's a quick-and-dirty prototype, not something that ought to be used in production or built on;

https://creativecommons.org/public-domain/pdm/

#AI #copyright #PublicDomainMark #PublicDomain

Blockchains - the tech behind "crypto" - are a tool for using computers to make abundant things scarce. There are only a tiny handful of ways - if any - to make destroying natural abundance a useful thing to do.

Backpropagation algorithms - the tech behind "AI" - are a tool for making reliable things unreliable. Computers themselves are *perfect* at math. But a Trained #MOLE will make mathematical mistakes, because what they do is statistically-weighted *guessing*, not computation.

(1/2)

#AI

Here's a challenge for any "AI" boosters out there;

Show us a benefit of using a Trained #MOLE that couldn't be gained, more or less, using auditable, programmatic automation, and at a much lower overall resource cost.

(1/?)

#AI

They're not "hallucinations", they're fictions. Every word that comes out of a Trained #MOLE is fiction, just like every image and video. When it mixes in elements of the real world, as all fiction does, that might make it more convincing, but it doesn't change what it is.

#AllAISlopIsFiction #AI

People keep comparing the "AI" bubble to the DotCom bubble. This is misleading.

What drove the DotCom bubble(s) was speculation on the transition from broadcast media to network media. Which was, and is, a decentralisation of media power. The emergence of #MOLE Training is exactly the opposite, a global re-centralisation of media power.

It's more comparable to the emergence of proprietary software, DataFarms and "the cloud";

https://www.gnu.org/philosophy/who-does-that-server-really-serve.html

#AI #DotComBubble #Speculation #TechHistory

@scottwilson
> They've announced an integration and partnership with ChatGPT

There was a time when sticking to Free Code apps was a good way to avoid hype-driven, fish-on-a-bicycle nonsense. Sadly, thanks to Open Source being so mainstream now (mostly a good thing), that hasn't been true since the crypto bubble.

Something about Trained #MOLE tech, in particular, seems to be irresistible to people who would usually know better, eg;

https://nextcloud.com/blog/ai-in-nextcloud-what-why-and-how/

#AI #OpenSource

Bwaaaaahahahahahaha! Is this guy living in another shard of the multiverse from me, or is this some kind of ironic, postmodern digital fiction product? Or the result of too much time conversing with a Trained #MOLE instead of human beings who read?

When it comes to vibe coding, one thing I haven't seen much commentary on here is the #UX. I know people with no prior software development experience, of any kind, who've turned their ideas into functional prototypes. Without having to learn the interface of a wireframing app, and try to wrestle their vision into wireframes (with all due respect to @penpot and their peers).

I know existing auto-coding tools come with *huge* issues. But That's not nothing.

(1/?)

#AI #MOLE #VibeCoding

"We’re building an AI that respects your privacy as much as it respects the planet."

How do you square this with the fact that all of the major #MOLE Training companies are DataFarming corporations, or create their models by leasing access to the huge volumes of data they hoard. Much of it personal data (collected from platforms they control), or arguably misappropriated (eg web scraping).

(5/?)

I just tried #Ecosia for the first time in years. It doesn't do anything without JavaScript, like Brave Search but unlike DDG, which has an HTML-only mode.

After letting Ecosia run JS, it tried to make me solve a ClownFlare CAPTCHA before letting me continue. Not impressed so far.

Like DDG and Brave, Ecosia is leaning into #MOLE Training, although it doesn't insert auto-summaries into its standard search results page;

https://blog.ecosia.org/what-we-are-doing-with-ai/

Search results are pretty relevant though.

(2/?)

#TIL the sad news that a major not-for-profit organisation for NaNoWriMo was a casualty of the #Mole Training craze, although apparently the writing challenge continues informally, much as it began (see #WritingMonth);

https://www.theguardian.com/books/2025/apr/02/creative-writing-nanowrimo-to-close-after-20-years

It shut down just over a year ago, after antagonizing their community with a public statement claiming that "the categorical condemnation of Artificial Intelligence has classist and ableist undertones";

https://www.404media.co/nanowrimo-ai-policy-classist-ableist/

#AI #NaNoWriMo #writing

Versión 1.49.2 de Mole, aplicación para macOS que limpia, desinstala, analiza, optimiza y monitoriza tu Mac desde la terminal: https://www.dekazeta.net/foro/files/file/4986-mole/

#Mole #macOS #Mac

I've taken to using Brave Search, in parallel with DuckDuckGo, as my general web search engine.

There's no obvious way to turn off the auto-summaries by whatever Trained #MOLE Brave are using, and unlike on DDG, they embed specific lists of standard search results within the text, in an attempt to back up specific points. So I end up kind of using it, even though I presume its output is nonsense until proven otherwise, which it often is.

(1/2)

#WebSearch #AI

#HatTip to the Te Aka whānau for a website that works without JavaScript, including a simple and very effective internal search. Perhaps there are extra functions that depend on JS, but I don't notice their absence.

https://maoridictionary.co.nz/

This is proof that you don't JS nor a Trained #MOLE to make functional, easy-to-navigate websites.

#MakeJavaScriptOptional #TeAka #TeReoMāori

@shtrom
> free advertising in any codebase.

I hadn't thought of that, but you're bang on. Some codebases use AGENTS.md, which at least avoids doing unpaid promotion for #MOLE brands.

It's akin to how DataFarming platforms get free advertising in shops ('find us on FarceBook'), free logos on other people's websites, and free mentions in podcasts ('review us on Spitify'). I frequently point this out when giving organisations feedback on their websites.

@mattcen @qdot

Kernel maintainers are among the heavy hitters of software engineering, responsible for the code at the core of the devices and services we all use every day. This is what one thinks about using a Trained #MOLE;

"Note, LLMs are very good at finding suspect 'security issues' in kernel code these days. But even with the best of the current and next generation tools, at least 1/3 of the results they generate are flat out wrong or harmful."

@gregkh, 2026

https://lore.kernel.org/all/2026080354-skater-urgent-31b2@gregkh/T/#u

#AI #LLM

We have a very shallow and lazy mole in our garden. #mole #gardening

To be fair, the Opportunity party policy on this is even worse. They're totally drinking the kool-aid, and planning to give free public money to the #MOLE Training and DataFarming industries;

"... we will introduce a new tax credit scheme.

It will allow claiming back as a tax/credit up to 25% of the cost of rolling/out high-tech solutions such as AI and digital platforms."

https://cdn.opportunity.org.nz/documents/policy/breakthrough-economy-policy-overview.pdf

This is the kind of corporate welfare Labour regularly disappoints me with.

#Opportunity

(2/?)

"Within that broader launch, the one substantive new policy announcement was a proposed one-year moratorium on new AI data centre consents."

#HenryOliver, 2026

https://thespinoff.co.nz/the-bulletin/27-07-2026/a-green-party-for-all-of-us

The Greens like to present themselves as radicals, but this is incrementalism writ large. To those of us closely following the disasterclasses of the #MOLE Training industry, this policy is about as ambitious as a 1 year moratorium on fossil fuel exploration, Or open field trials of GM.

(1/?)

#AI #datacentres

@bluewave_kiwi
> tiktok is a Chinese information and propaganda operation

No more so than any other DataFarming social media platform. Particularly now that most of isn't owned by a Chinese parent company.

The majority owners are now;

Silver Lake: a private equity firm who sold Skype to BorgSoft

MGX: an entity owned by the UAE, set up to funnel money into #MOLE Training

Oracle: 'nuff said.

So if anything, it's now a *US* information and propaganda operation.

#TikTok #Oracle

"'Checking out a parallel copy of our Go repository and telling the AI to rewrite the whole thing in Zig while I work on something else just so I can keep my job. I hate this shit so much. My job has usage tracking and quotas. I don’t use it for actual work, I just spin it up and disregard the output'.

– An actual software engineer"

#NikhilSuresh, 2026

https://ludic.mataroa.blog/blog/ai-mania-is-eviscerating-global-decision-making/

So apparently much of the measurable demand for a Trained #MOLE services isn't actually real ...

(2/?)

#AIBubble

Я стесняюсь #humor #music #mole #story #shorts

According to This Machine Kills on 12 Jan, Bernie Sanders put forward a bill proposing thar the US federal government buy shares in the massively overvalued companies inflating the "#AI" bubble. The bubble that Michael Burry has already taken out a number of shorts on (see The Big Short). Buying into that, seems ... profoundly unwise.

Someone needs to take Uncle Bernie aside, and explain that the MOLEs aren't smart, and the #MOLE Trainers are running an elaborate, large-scale financial scam.