m-hodges

6 ms·

91.
by m-hodges 5mo ago
There is literally zero mention, or even indication, of “election fraud” in this article other than the headline.
92.
by m-hodges 5mo ago
> Agent needs an inbox and hits AgentMail via curl. Agent receives instructions via MD I'm fairly AI-optimistic, but I feel like I'm taking crazy pills. Every day the HN story is either "Apple patches actively exploited ze
93.
by m-hodges 5mo ago
what is a "professional pc builder" in 2026
94.
by m-hodges 5mo ago
To the “LLMs just interpolate their training data” crowd: Ayer, and in a different way early Wittgenstein, held that mathematical truths don’t report new facts about the world. Proofs unfold what is already implicit in axioms, definitions,
95.
OpenAI Prepares to File to Go Public in Coming Weeks(nytimes.com)
6 points by m-hodges 5mo ago|0 comments
96.
Show HN: LLM Steganography to hide text inside another plausible text(github.com)
2 points by m-hodges 5mo ago|0 comments
97.
US Power Investment Is at Record Highs, Driven by a Solar and Battery Boom(apricitas.io)
2 points by m-hodges 6mo ago|0 comments
98.
LLMs can hide text in other text of the same length(arxiv.org)
5 points by m-hodges 6mo ago|0 comments
99.
Our Future Is Being Devoured by Feral Thought Experiments(programmablemutter.com)
5 points by m-hodges 6mo ago|0 comments
100.
AI's biggest critic has lost the plot(theargumentmag.com)
40 points by m-hodges 6mo ago|50 comments
101.
by m-hodges 6mo ago
> Keep that in mind whenever you read breathless headlines about AI causing layoffs or dampening hiring. It’s something like 1% of all firms - split roughly evenly between “increasing employment due to AI” and “decreasing employment due
102.
AI Adoption in the United States and Its (Tiny) Labor Market Impact(macromostly.substack.com)
2 points by m-hodges 6mo ago|1 comments
103.
Anthropic has surged to a trillion-dollar valuation, overtaking OpenAI(businessinsider.com)
4 points by m-hodges 6mo ago|0 comments
104.
by m-hodges 6mo ago
Excited for OpenClaw to email Workspace Agents.
105.
by m-hodges 6mo ago
Periodic reminder that you can just buy a cheap IC¹ to rip GPS data signals right out of the air. I built a GPS-powered clock that sits on my desk. ¹ https://www.sparkfun.com/gps-module-gp1818mk-56-channel.html
106.
UK Permanently Bans Tobacco for Anyone Born in 2009 or After(nytimes.com)
3 points by m-hodges 6mo ago|1 comments
107.
Nodes, Edges, and the Historian's Craft: Knowledge Graphs as Research Notes(computationalhistory.substack.com)
2 points by m-hodges 6mo ago|0 comments
108.
by m-hodges 6mo ago
"But though our thought seems to possess this unbounded liberty, we shall find, upon a nearer examination, that it is really confined within very narrow limits, and that all this creative power of the mind amounts to no more than the
109.
Artful Mathematics: The Heritage of M. C. Escher [pdf](pub.math.leidenuniv.nl)
3 points by m-hodges 6mo ago|0 comments
110.
Eschers Print Gallery and the Droste Effect(mathvisuals.org)
2 points by m-hodges 6mo ago|0 comments
111.
by m-hodges 6mo ago
The author also noted: > yes this has to buy below 0.73 long term, the bot has a configurable ceiling set at 0.65 and checks for new markets buying closer to .5 https://x.com/sterlingcrispin/status/2043685362812
112.
by m-hodges 6mo ago
Not my project, but author said on X: > Why predict the future when 73.4% of all Polymarkets resolve as No? https://x.com/sterlingcrispin/status/2043398710013595857
113.
Closing the AI Open-Source Gap(newsletter.ownersnotrenters.com)
1 points by m-hodges 6mo ago|0 comments
114.
Nothing Ever Happens: Polymarket bot that always buys No on non-sports markets(github.com)
471 points by m-hodges 6mo ago|276 comments
115.
by m-hodges 6mo ago
I tried it a few times and always found it disappointing. It typically started off like a structured "lesson" but as I chatted with it, it would forget the syllabus is had proposed and we never "completed" the thing we s
116.
Small Models Are Smart Enough(twitter.com)
18 points by m-hodges 6mo ago|1 comments
117.
Surviving Hallucinations: AI Fault Tolerance for Campaigns(matthodges.com)
3 points by m-hodges 6mo ago|0 comments
118.
Treasury Secretary and Fed Chair Convene Bank CEOs about Mythos Model Risks(reuters.com)
6 points by m-hodges 6mo ago|0 comments
119.
by m-hodges 6mo ago
I think this was a pretty honest write up of what went well and what didn't, and I think it's directionally pragmatic on takeaways. > The agent knew the experiment ended at Day 30 since I told it as much in the system instructi
120.
"It's hard to feel like Claude isn't actively working against me."(twitter.com)
3 points by m-hodges 7mo ago|1 comments

More ›