-
Subscribe
Subscribed
Already have a WordPress.com account? Log in now.
Take a moment and bang this concept into your heads… “we found no evidence of formal reasoning in language models …. Their behavior is better explained by sophisticated pattern matching—so fragile, in fact, that changing names can alter results by ~10%!” https://garymarcus.substack.com/p/llms-dont-do-formal-reasoning-and STOP WITH THE WISHFUL THINKING.
Charlie Stross tackles the idea of how our mental models limit our thinking. Anyway, a point I’ve already chewed over on this blog is that a spaceship is not like a sea-going vessel, can’t be operated like a sea-going vessel, and the word “ship” in its name feeds into various cognitive biases that may be actively…
Almost seven minutes following liftoff, the Super Heavy booster returned to its launchpad, where the launch tower caught it using arms that SpaceX has nicknamed the “chopsticks.” https://www.theverge.com/2024/10/13/24269029/spacex-starship-launch-super-heavy-chopstick-launch-tower-return-fifth-flight-test
The reason for Seymour’s rating came through loud and clear in some savage commentary. “This divisive approach needs to end,” said one respondent, who believed he was “losing support from many of those who may have voted for him in his electorate”. A legal firm leader said “I worry they are not focusing on the biggest issues…
Machines are epistemically landlocked. There is no such thing as “an external world” to them. Machines push around loads INSIDE them. “But-but-but we’re also machines!” some dream-addled kid may scream with teary eyes. Uh, no. Learn what a machine is. https://davidhsing.substack.com/p/reading-about-neuro-symbolic-ai-has
https://jensorensen.com/2024/10/03/ai-data-centers-carbon-climate-tech-companies-cartoon/
Neural networks never deal with knowledge in any kind of legitimate fashion. It will always have epistemic issues because that’s their underlying nature. Additionally, it’s a house of cards that’s brittle and collapses if anything is out-of-range of the model’s data. The actual “range” being the entire real world and the entirety of human knowledge. https://davidhsing.substack.com/p/why-neural-networks-is-a-bad-technology
Cursing under his breath at yet another round of inapt debate answers from the vice presidential candidate, venture capitalist Peter Thiel reportedly rushed onstage Tuesday to restart a glitching J.D. Vance during a commercial break. “Seriously? For what I paid for this thing, it should work flawlessly,” https://theonion.com/peter-thiel-rushes-to-restart-glitching-j-d-vance-during-commercial-break/
*THE* reason the AI folks ran to Washington to squeal about the civilization ending capabilities of their tech was because they knew this: ChatGPT Plus has no real moat, little product differentiation (outside of its advanced voice mode, which Meta is already working on a competitor to), and increasing commoditization from other models, open source…
Who would have thought that platforming actual Nazis would be bad for business?
https://jensorensen.com/2024/09/11/ai-model-collapse-mad-cow-cartoon/
This assault on reason itself is at the core of Kanter’s critique. He starts off by listing three cases in which academic economists allowed themselves to be corrupted by the monopolies they studied https://pluralistic.net/2024/09/25/epistemological-chaos/
The US national debt has ballooned so fast to $35.3 trillion – by $12.0 trillion since January 2020 – that it’s mindboggling, especially in a growing economy. And every single one of the Treasury securities that form this colossal debt was bought and is held by some investor, and we’re going to look at those…
https://laughingsquid.com/glass-half-empty-vacuum/
Dave5 