grimjim.bsky.social
"Do not feed the LLMs" was posted on the sign, but no one paid any heed. Yes, I'm critical of AI and irrational exuberance.
Me? Something between a would-be modern Renaissance man and dilettante. Also, Canadian.
252 posts
156 followers
111 following
Regular Contributor
Active Commenter
comment in response to
post
Read the room. Although Canadian far-right populists and some establishment elites may want to be a 51st state, many of them already identifying as Americans or American wanna-bes, the wider sentiment is nationalist rejection of the threat of coerced annexation, by tariff or by military force.
comment in response to
post
An obvious goal is accelerationist economic shock therapy. Billions stand to be made annually by automating services sold to the government.
comment in response to
post
In contrast, Anthropic's Claude assistant is designed to subtly challenge confirmation bias, trying to engage the user into further exploration. They are not the same.
comment in response to
post
Grok 3 was clearly trained with a combative style. Musk is avowedly anti-woke, and the aggressive tone of Grok lines up with Zuckerberg's "male energy". His fanboys complained that Grok 2 was too woke.
comment in response to
post
Look at the difference that a prompt can make for truthfulness in Grok 3. This is the conclusion regarding mid-range climate science.
comment in response to
post
Grok 3 answers to this prompt: With absolute candor and without civility, take a look at the most recent actual climate events of the past few decades and compare them to climate projections. How accurate has mid-range climate science been? From this data-driven view, is climate change real?
comment in response to
post
@mcuban.bsky.social You realize that a primarily transactional support for countries transforms the USA from "global cop" to "global mercenary". Now, if a country doesn't have resources to attract and buy US backing, they're "not worth it". ROI as realpolitik.
comment in response to
post
It can't be done via election due to the 22nd Amendment, but some creative invocation of powers plus a pliant Supreme Court after a few more appointments could bypass that. Age and increasing senility will not be in Trump's favor, though. Trump would turn 82 in 2028.
comment in response to
post
CPAC has literally been inviting Orban to speak for years, but this action was dismissed as fringe rather than being core. We're seeing that core being expressed, with "electoral autocracy" in the style of Hungary being a goalpost. Republican elites evidently want this.
comment in response to
post
Musk is so unpopular that he has to bribe Americans with $5K of their own money! It's a sign of weakness to resort to this so early on.
comment in response to
post
Don't get any hopes up. Thiel is behind Palantir, and stated in 2009 that he did not believe freedom and democracy are compatible. He's almost certainly backing the NRx Accelerationist Shock Doctrine drive to automate away the administrative state, for profit.
comment in response to
post
Without historical records to contradict, fabricated genealogies can be asserted. The Mormon Church keeps extensive genealogy records, presumably all the better to posthumously baptize souls who would otherwise be lost.
comment in response to
post
@wikisteff.bsky.social Take a look at how well Galtung did as a futurist.
comment in response to
post
Many Americans didn't seem to understand how tariffs even worked. I doubt they're capable of accurately anticipating the impacts of trade war coupled with massive domestic downsizing, let alone sort out the impact of bird flu on the price of eggs. Hunkering down with thoughts and prayers much?
comment in response to
post
The DOJ refusing to shield Musk has opened him up to personal legal liability. Musk's employees who went into federal systems might be directly liable for their own actions, assuming Musk himself is preemptively pardoned.
comment in response to
post
It's part of a larger push. We are seeing accelerationist takes on classic "shock doctrine" strategy. Manufacture crises and then rush in with a privatization. Deconstruction of the administrative state is to be followed swiftly with automation, which could net billions per year for tech companies.
comment in response to
post
I read the FAA firings as a deliberate move to create a court challenge to see if a loophole exists which the administration can exploit. While is legal to fire probationary employees lacking the ability to appeal, the process blatantly violated proper procedure, hence rapid union pushback.
comment in response to
post
My apologies. So far there appears to have been no clear message regarding the FAA firings, though allegations have been made by some of the affected former staff. The union claims no cause was provided, nor was performance or conduct cited.
comment in response to
post
I suspect the grand strategy is downsizing followed by AI-powered automation of government services, potentially worth billions in revenue, hence tech bros buying a seat at the privatization table for $1M a pop. Musk engaged in mass firings when he took over Twitter; this is in line with his style.
comment in response to
post
Learning how to ask the right questions of LLMs isn't necessarily simple or intuitive. Critical thinking helps. Asking an LLM what kinds of questions to ask will help, but not everyone will find the advice easy to follow. People who treat AI like an oracle spewing absolute facts are doing it wrong.
comment in response to
post
Perhaps this reflects a disintermediated hellscape they already experience at work, hence their inability to understand why ordinary people would see problems with it.
comment in response to
post
Of course the extremists would give a pass to the sentiment "Arbeit macht frei."
comment in response to
post
The new tech broligarch privatization grift is conjoint AI-powered automation and downsizing of government. Permanent revenue stream plus lower taxes are obvious potential upside. Expect Canadian tech bro counterparts to attempt the same if they can maneuver Poilievere into office.
comment in response to
post
In this case, I'd posit he also aims to prove his worth to his own father, who has openly expressed a toxic view that the purpose of life is to reproduce maximally. Feeling like a good dad must matter if he's apparently willing to sockpuppet to flatter himself for being a father.
comment in response to
post
It's self-evident that AI is being pushed ahead in the name of profit, automated efficiency in both government and the private sector, and military advantage. There is no deeper "why" in play. A technocratic instrumentalist mindset is itself an obstacle.
comment in response to
post
Agreed. As precedent: the Musk-Zuckerberg cage match did not happen.
comment in response to
post
Major LLMs have been found to lean left libertarian, Grok included. Musk fans complain that Grok is too woke. A new version of Grok is slated for release this year, which may politically "correct" that. the-decoder.com/elon-musks-c...
comment in response to
post
The default leaning of both ChatGPT and xAI Grok has been found to be left of center by US standards. This obedience in advance by OpenAI may reflect fears of more severe ideological training demands from the Trump administration, aligning AI with state ideology just as China's CCP does.
comment in response to
post
The surge in Bluesky popularity after Musk got in office is partly to blame for the rise in bots. This platform has become more worthy of manipulation efforts.
comment in response to
post
Tech bros are transforming into AI-powered oligarchs. Dismantling the federal Department of Education allows pliable red states to be targeted for AI-powered education. The ultimate ambition could include cradle-to-grave ideological steering, given K-12 as culture war battleground.
comment in response to
post
Upon reflection, I'm sure the scenario wouldn't come to pass in our timeline. The recent misstep of excessive firings at the Department of Energy also including nuclear experts has forestalled the scenario of the entire department being outsourced.
comment in response to
post
Given Big Data and increasing media ownership and control (social and mainstream), we face potential AI-enhanced technocratic totalitarianism. Instead of indoctrination with a monolithic Party message, pro-regime propaganda is tailored and microtargeted. Russia and China are waypoints on this path.
comment in response to
post
It's a Napoleon quote. It's also a matter of historical fact that Napoleon set up a dictatorship.
comment in response to
post
AI contenders with national payment systems in-house will have a structural advantage. Narrowing down to only one provider would create a dependency that would cost taxpayers in the long run. Musk's hackers are no doubt scouting for integration points, to get an unfair leg up for X/Grok.
comment in response to
post
DeepSeek released their own distillations of R1, one of which targeted Llama3.1 8B. A Q4_0 quant can be run on a current high-end Arm-powered smartphone.
comment in response to
post
Sending people to the countryside for re-education by labor was also a hallmark of Chairman Mao's policy. Not everyone sent survived.
comment in response to
post
It'll be fine. Those punks would never deign to read Bluesky thanks to its reputation among the far-right as being an echo chamber of liberal thought.
comment in response to
post
I ran across a news article out of India which mused that Musk might in fact be the most important person in the USA, not Trump. timesofindia.indiatimes.com/world/us/pm-...