neuronexmachina

Whenever I hear about this I can't help but be reminded of the Howard Hughes Glomar Explorer, where a hugely-expensive commercial project was actually part of a secret CIA initiative to recover a Soviet submarine: https://en.wikipedia.org/wiki/Glomar_Explorer

I'm not sure that's truly the case here, but it does seem like the tech Project Suncatcher is working on has at least some overlap with requirements for military SIGINT and in-orbit imagery processing.

It's kind of crazy that claude-opus-5.5 (max) uses 119k tokens per "intelligence index task" while gpt-6-astra (max) uses 27k.

https://artificialanalysis.ai/models?cost=intelligence-vs-co...

> The incident of US almost boarding a Chinese ship due to an AI-assisted intelligent report claiming it was transporting nuclear weapons components[0] was more compelling for me in terms of direct AI usage impact.

To be fair, there was a rather similar incident in the 1990s, so that sort of thing can happen without any AI at all:

https://en.wikipedia.org/wiki/Yinhe_incident

I was curious about what various forks actually do, found a list of some here: https://commaguide.com/compare/

They've trademarked it for quite a few closely-related domains: https://tsdr.uspto.gov/#caseNumber=87709072&caseSearchType=U...

>Computer software, namely, game engine software for video game development and operation; Computer software, namely, software development tools for the creation of computer-generated imagery and graphics for the production of video games; Computer software, namely, software development tools for the creation of computer-generated imagery and graphics for the production of content for virtual worlds and 3D platforms; Computer software, namely, software development tools for the creation of computer-generated imagery and graphics for the production of motion pictures, television shows, videos, 3D animations, 3D simulations, 3D visualizations, virtual reality motion pictures, virtual reality television shows; Computer software, namely, software development tools for the creation of computer-generated imagery and graphics for the production of virtual reality video games; Virtual reality game software; Virtual reality software for creating multimedia content; Augmented reality game software; Augmented reality software for use in mobile devices for integrating electronic data with real world environments for the purposes of entertainment

Yeah, I was hoping it would be something to make it easier for agents to interface with Unreal Engine games.

Enterprise and most Team accounts use API pricing, they don't have an included-usage quota.

A recent example was the Google workspace CLI, which was maintained for just a month: https://github.com/googleworkspace/cli

What would you consider as sufficient evidence of LLMs being useful in a particular domain?

I mean, they kind of do? I assume they're hesitant to write a how-to on how to tailor the exploit image:

>These are not out-of-the-box exploits. Exploitation requires fingerprinting the target version and tailoring the payload image(s). Some of our RCE attempts landed only after thousands of image uploads. That said, an AI agentic approach with a frontier model like GPT-5.6 Sol cut exploit development time down to roughly 1 to 3 days from initial probe to remote RCE. A motivated attacker can convert a vulnerable upload endpoint into RCE or an info leak.

I'm not sure, but Sir Urquhart may have invented the word, and he's essentially the only person to have ever used it. It's presumably based on the latin quirites: https://en.wikipedia.org/wiki/Quirites

> 1. they apparently gave advice to both the clinton and trump campaign

Just because this is worded ambiguously, to clarify the "they" is Facebook. CA on the other hand was effectively the Trump's digital ad campaign.

> 2. the trump campaign did targeted marketing better (ie. targeting the base rather than the marginal voter)

This ignores that a lot of what made CA different is that they used microtargeting to discourage turnout via fake news stories.

3. "I'm skeptical that Cambridge Analytica has any data or expertise that couldn't be replicated readily by someone else"

After CA's methods and actions became public, FB massively limited what you could do via their API.

> We should be outraged and OpenAI/Anthropic should be (and in my mind, are) legally liable for the crimes they've committed thus far.

How would that work out if these were self-hosted open weight models?

It's worth noting that Equal Earth isn't an equidistant projection. The upper bound on distance distortion is an order of magnitude less than Mercator's though.

Counterpoint: the Cold War was pretty much all about geopolitical symbols, e.g. the Berlin Wall, nuclear deterrence, the Space Race, counting Olympic medals.

> I know the UN claims that "Correcting cartographic distortions of continental sizes constitutes an act of cognitive justice and memorial reparation." but it's purely symbolic and changes nothing geopolitically.

This is kind of personally embarrassing, but due to seeing Mercator maps all the time growing up it wasn't until later in life that I realized that Greenland isn't actually a gigantic landmass the size of Africa.

> The value of a geographic projection is in its utility for the task at hand

Other than maritime navigation, I can't think of many tasks where Mercator would be more useful than an equal-area projection.

There's a separate post about the technical differences: https://cursor.com/blog/git-at-any-scale

Is this basically trying to do something similar to Zed's https://delta.dev/, but from Slack instead of an IDE?

It's pretty long, but looking through it the text seems pretty streamlined. I think there's just a lot of changes.

Reversing the test order is also pretty effective: https://pypi.org/project/pytest-reverse/

I imagine Codeberg's LLM ban might make it a poor fit for many Cursor users.

Zed releases new stable versions weekly, with plenty of features and bugfixes in each version: https://zed.dev/releases/stable

From one of the linked sources: https://womenhack.com/women-in-tech-statistics/

>Women leave tech at 45% higher rates than men. [Accenture, 2024]

That doesn't seem to be true: https://www.cio.com/article/201905/women-in-tech-statistics-...

>Women leave the tech industry at 45% higher rates than men, and 50% of women leave the tech industry by age 35 compared to only 20% for other industries, according to Accenture.

If I understand correctly, for speed of deployment they're building a less-efficient simple-cycle plant (instead of the typical combined-cycle). Even though it's larger, the efficiency is worse than a typical natural gas plant.

Whose security is intentionally crappy in this scenario?

Cities shouldn't fund the software they rely on?

Chrome uses libexpat for XML parsing, not XSLT.

It does mention it though?

> Many robotic applications need to operate without network latency or internet connectivity. Gemini Robotics On-Device 2 is built specifically to handle these constraints — it is our most-efficient vision-language-action model (VLA) optimized to run locally on robotic devices.

> If that is really what they are concerned with, they probably could have just had Claude Code rewrite Claude Code in rust

Rewriting a runtime with well-defined interfaces and behavior is quite different from rewriting a user-facing application under active development by a decent chunk of their organization.