@wwhitlow@indieweb.social

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

Replying to

@derickr@phpc.social @elduvelle@neuromatch.social @toni@zug.network @Crell@phpc.social What can be said for StackOverflow, it is certain that the landscape of knowledge is changing. Looking to Wikipedia is probably the best approach for the AI free writing policy. https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing This new article from Anthropic might be of interest. It test skill acquisition of Junior Devs with and without AI. The sample size is somewhat small, but may offer some insight. https://www.anthropic.com/research/AI-assistance-coding-skills

en.wikipedia.org

5

0

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

So Ring just admitted that they’re scanning everyone’s video feed with #AI systems. How is that not considered an absolute security vulnerability?

#superbowl #ring

indieweb.social

Indieweb.Social

2

0

2

0

William Whitlow @wwhitlow@indieweb.social

· 7mo ago

Interesting paper that was recently updated today engaging Heidegger's philosophy with contemporary Machine Learning techniques. I hope to take more time to engage with this paper over the next couple of days. Still, it is encouraging to see such consideration connecting AGI concerns with philosophical principles. Exploring how contemporary design principles lead more to tool use, than AGI.
https://arxiv.org/abs/2602.19028

#AI #ML #AGI #philosophy

arXiv.org

The Metaphysics We Train: A Heideggerian Reading of Machine Learning

This paper offers a phenomenological reading of contemporary machine learning through Heideggerian concepts, aimed at enriching practitioners' reflexive understanding of their own practice. We argue that this philosophical lens reveals three insights invisible to purely technical analysis. First, the algorithmic Entwurf (projection) is distinctive in being automated, opaque, and emergent--a metaphysics that operates without explicit articulation or debate, crystallizing implicitly through gradie

1

0

1

0

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

Replying to

@UlrikeHahn Agreed. Before seeing your post, I rather assumed there wasn’t anywhere near this level of moderation. I suppose the proliferation of AI research centralizing on this pre-print resource has led to this change.

1

4

0

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

Replying to

@Crell@phpc.social Thank you for quantifying the various aspects that it takes to keep AI running. It’s one of those details that is becoming increasingly apparent, but difficult to grasp the scope. Would you care to elaborate on the “This is how societies die.” comment? Is that primarily in the sense of social apathy? In the lack of respect for others/foresight over long term consequences? The disruptive tactics of numerous tech companies that have eliminated many norms? Or something else?

1

23

0

William Whitlow @wwhitlow@indieweb.social

· 7mo ago

Regardless of expectations for AI systems, interpretability studies seem to be promising towards the future of understanding the mathematical associations of concepts. Breaking down the model to its most atomic representation. This paper explores some of the associations regarding trust. It would be interesting to see if there's a correlation between the embedding of human trust models, the persona vectors of various models, and the ability to jailbreak.

https://arxiv.org/abs/2603.05839

#AI #LLM

arXiv.org

Evaluating LLM Alignment With Human Trust Models

Trust plays a pivotal role in enabling effective cooperation, reducing uncertainty, and guiding decision-making in both human interactions and multi-agent systems. Although it is significant, there is limited understanding of how large language models (LLMs) internally conceptualize and reason about trust. This work presents a white-box analysis of trust representation in EleutherAI/gpt-j-6B, using contrastive prompting to generate embedding vectors within the activation space of the LLM for dia

0

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

So ai.com is currently getting the hug of death from #superbowl traffic. However, simple search result seems similar to moltbot, clawdbot, or moltbook. Given all the security problems these have produced, Super Bowl publicity and concerning. Does anyone else have more information?

#ai

indieweb.social

Indieweb.Social

0

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

Replying to

@strypey @UlrikeHahn Wow! I wonder how much effect notions like Goodhart’s law is having on this. As individuals focus on publishing to increase the credibility. Unfortunate that AI misinformation in papers has seemed to proven the need for this to some degree. Still the clause that appeal is only available if the paper has been approved for publication elsewhere. That’s a final decision that offloads moderation to another team.

0

10

0

William Whitlow @wwhitlow@indieweb.social

· 8mo ago

Replying to

@UlrikeHahn I didn’t take that to mean that you blamed arXiv for this issue. My comment was more so to point out the difficulty in proposing any solution. As such an appeal process indicates that arXiv has made their final decision and will only change as a result of outside feedback. The gamification of systems makes me wonder if this would lead people to find a simple peer review journal in order to force the appeal. Simple thoughts on a complex issue.

0

8

0