Ideally, the prosecutor bears the burden of proof. We generally shouldn't impose systems that require defendants to prove a negative. I recognize that reality does not necessarily match this ideal.
It's ultimately up to juries to decide whether a defendant's assertion that evidence is fake is enough to constitute reasonable doubt in the absence of hard evidence for it. I imagine that's going to be very context-dependent. It would probably work if I was accused of this, with no history of anything like this, versus a guy who does this frequently, posts videos of himself doing it regularly, and never gave any indication they're fake until he got in trouble.
M3 Ultra with 256 GB memory, using GPT-OSS 120b in ollama. It’s decently fast, but makes the system somewhat unstable. Have to reboot frequently otherwise the GPU seems to flake (eg visual artifacts / glitches in other programs).
This reminds me of Snake Galaxy. I'm so sad it's not available anymore, it was an adorable spherical snake game on early iPhone circa 2010 era. I really miss the little Paris planet.
My first hand experience with the deterioration of quality in MacOS: I bought a Mac Studio to run some traditional ML neural network type stuff with PyTorch. Unfortunately the GPU / MPS kept crashing (?!) and would require a full reboot of the computer to get working again. Without a reboot, PyTorch would not find a GPU / MPS device and random other apps would glitch. After finding nothing in logs and no information about how to debug anything related to MPS, I contacted Apple Support. Apple support was completely unhelpful and a very frustrating experience. It felt like they were laundering responsibility through the veneer of some engineer having glanced at whatever internal second hand report the support staff transcribed from me. The situation improved a bit after upgrading to Metal 4, but in the end I moved back to Tensorflow. It's very disappointing to spend $6k on a machine and the software doesn't work as expected.
People tend to assume things would go according to their fantasy mental model on this topic, but I strongly believe that legalisation of strongly addictive substances would pave the way for an illegal market for stronger and cheaper illegal drugs, untaxed and unregulated.
Addicted people don’t reason the same way as non-addicted people.
Disagree. Do you mean that vehicles gets louder because people skip maintenance? Or more stress produces more honking? An opposite effect: more financial stress means less spending on aftermarket exhaust mods means quieter traffic. Also, financial stress means less driving means less noise. Overall, I doubt you're going to find much signal in the noise!
Well, it is a perfect storm of factors. First of all lack of maintenance. Secondly, people gunning it with the gas pedal. Thirdly, lower income folks in a rush to get to a second or even third job, gunning it for reasons other than just sheer frustration.
>We test three representative tasks in materials chemistry: linking dopants and host materials, cataloging metal-organic frameworks, and general composition/phase/morphology/application information extraction. Records are extracted from single sentences or entire paragraphs, and the output can be returned as simple English sentences or a more structured format such as a list of JSON objects. This approach represents a simple, accessible, and highly flexible route to obtaining large databases of structured specialized scientific knowledge extracted from research papers.
Short answer: It’s a way to generate structured databases for (most) scientific topics. Why? Apply data driven methods to these databases. So what? It’s a powerful way to ask and investigate scientific questions/trends otherwise hidden inside a million scientific papers.
Example: Consider what PDB has done for our understanding of protein folding, as well as the ML/computational techniques they’ve enabled (eg, Alphafold). Most scientific questions and properties are not as data-rich as protein folding. What if they could be?
Longer answer: The last 15 years in computational/ML + science have shown that structured databases open up entirely new frontiers in discovery (eg Protein Data Bank, Materials Project). But most scientific topics/properties are NOT in structured DBs, they’re scattered about in millions of papers. It’s especially a huge problem in some topics in materials science. It’s not that these problems are data scarce, but that it’s hard to actually collate their data in a structured format. You literally cannot use most ML methods because structured DBs do not exist.
This paper is a way to generate massive structured databases of specialized, intricate, and hierarchical knowledge graphs from scientific literature. Fine tuning works, prompt engineering does not (at the time, perhaps this has changed). Once you have a database, you can analyze an entire subfield or topic in science with ML or stats methods.
