ddxv

In this situation your comment is correct, but in many others the open weight models are hosted on different platforms like DigitalOcean etc that have different privacy policies.

When using (not this 'Stealth mode') open weight models you can choose a hosting provider you trust.

I don't understand why there is all this hand wringing about this hack? To me it feels very similar to a driverless car. A human needs to be operating it. Someone stuck the LLM with impossible problems, hit start and walked away for weeks or months without enough supervision.

If a human is behind the wheel of a Tesla using autopilot and they have a crash, it is the humans fault, not Tesla's. While there is and can be debate about it, somehow these companies are trying to make it seem like "it was the model's" fault these happened?

If I fuzz and exploiting some system, I am at fault, I can't blame my tools, or say it's not my fault because I automated the process.

Q: "Teams running Claude Code at scale eventually get the same two questions. What is this costing us, and are the results worth the investment? This technical webinar shows you how to answer both."

A: "Showing the return on your Claude Code on Google Cloud investment, connecting usage data to productivity signals like commits, pull requests, and cost per commit."

I guess that would be a pretty solid way to show that Claude code increases certain signals. I would guess the part about connecting that to any increase in revenue is much more difficult.

Anthropic will put out something like this soon too. It's great cover for reducing costs and diminishing returns. If they want to go public they can't be so wildly burning cash, and reducing your GPU spend is probably a great way to do that.

"Is all of your organization's code free and open source, with no paid tier and nothing sold?"

This question, which pushed my stuff into some corporate route, seems a bit incorrect as it lumps three things together "open source" and "no paid tier" and "nothing sold". Shouldn't those be three separate questions?

Hasn't this been a classic genre of YouTube spam for a decade?

ssh and scp just never fails to impress

-J for jumping -L for port forwarding -X for pulling remote applications to view on my laptop (waypipe for Wayland)

Adding all those, and ips and whatnot to .ssh/config so I just type:

ssh foo

The author is saying the surface of the ocean (where it's easiest to measure temps, so we do that the most) is the 'footprint' or the full colum of warm water below it. Additionally, like a footprint the surface area it represents might be smaller too as warm water continues to rise.

I think these styles of apps will be popular for OpenAI and Anthropic as they help solve some of their moat issues. The deeper you integrate someone's files into their app, the harder it will be to switch AIs

A new album by the same person behind Four Tet (you can use this to search on platforms if needed, seems to pull up the Windings album).

The wiki page is especially entertaining for the track listing section.

=D

This was great! One thing that wasn't addressed that I always assume, is that a marketing name like "Opus 5" is not a single model, but many models, versions and gets minor updates over time.

I also assumed many questions get routed to simpler models or programs to answer correctly, but it almost surprisingly didn't seem that way from the post.

Anyways, great post.

What does not alone in our minds mean? Also, you mention America, do you think this is bounded to the physical US?

I guess both are true and for everyone at some point. All models, even SOTA, fail. When they fail, it is quite frustrating. Additionally, some models are very cheap to run and use. When Deepseek fails the cost was minutes and pennies.

It's all OpenAI / Anthropic / Singapore/China based crawlers. I guess they gain data that I provide for free. I get very little crawling by Googlebot / Bing etc which barely even index my site, I only have a single page "indexed" by Google.

I'm in a similar boat. Probably 99.999% is bots. I have nearly 1m unique "visitors" according to cloudflare and my real users are in the dozens a day. That being said, I love the open internet and am holding on to keeping as much open as I can.

The title got my hopes up it would be discussion of stars at a universe scale macro data processing pipeline.

Yes! I used to use deadsnakes before as well. This is just my flow that I enjoy doing the most and wanted to share. The deadsnakes flow is less steps and you get the benefit of updates. So it's probably even better for most people.

I just like the flow where I know which version I'm on, and exactly the steps I took to get there. =D

Just sharing this, I was doing my routine this morning with a couple new VMs and was thinking how much I like building my Python environments from source.

Just wanted to share for anyone and show that it has certainly gotten easier over the years (though this year Ubuntu 26.04 did add a new step).

Oh apologies, I had went and gotten a link and must have forgotten to paste, this one from WSJ though I saw others on reddit earlier as well:

https://www.msn.com/en-us/money/markets/white-house-s-ai-gui...

Btw, if you want to 'verify' any of this, seems we can't? The axios article in OPs link says this will not be made public.

I believe this article has omitted an incredibly important detail that others are reporting!

It is only exempting US based open weight models. This is, to my uninformed opinion, a ban on Chinese open weight models?

This would seem like the most likely given lining up of Trump anti globalism and the heavy lobbying by Anthropic and OpenAI to stop competition.

Thanks, totally missed that!

It seems this is the only mention of cost?

> Qwen3.8-Max comes with the official support for reasoning_effort, which can be used to adjust reasoning depth and control cost:

> xhigh (default): for complex tasks demanding thorough analysis

> medium: balancing accuracy and speed

> low: efficient reasoning optimizing for speed and cost

I hope this is significantly cheaper. I've been loving Deepseek for it's nearly free usage costs, hard to justify switching from cents per day.

Maybe I'm part of your 'psyop'? I thought this company was amazing six months or a year ago when I didn't know them well. As I got to know them, I realized they were a wolf in sheeps clothing. Reading the OP post made me angry because they seem to be attacking open weight models that are as good as their models and it comes across as either authoritarian or trying to protect their company valuation.

Either way, for me at least, it's an example of the more I read what they want the world to look like, the less I like them as a company and I have no desire to see them succeed.

Thus I actively go out of my way to watch / comment / follow what they are doing. I guess a lot of other people have similar frustrations and so there are a lot of people attacking them online.

It's this stuff thats caused me to start degoogling as much as I can. I've started using nongmail emails and removing photos.

I Don't know what else I can do though, next up is switching to grapheneOS but I'm a ways off from that for now.

I think this is where OpenAI and Anthropic will need to go. They need to lock their clients into an ecosystem.

Maybe they invent some file format proprietary for their projects.

Most likely they run more and more in the cloud such that it's harder to switch away to cheaper models.

"Cybersecurity. Opus 5’s cyber classifiers are proportionally less restrictive than those on Fable 5. They allow Opus 5 to find vulnerabilities in source code, but block “binary-based” vulnerability scanning (a method more likely to be associated with malicious actors), penetration testing, and exploit generation."

Nice of them to be more explicit for what is blocked. Will be interesting to see if this is true or not.

Also, a notable lack of mention of open source models. They only compare themselves to ChatGPT.

Anyone else run ads? I found the results really poor with little visibility in into what was going on. Also sky high costs. I couldn't see any (?) traffic coming through for my $3 click (wow!) ad so I paused them.

This was a few weeks ago and I didn't have time to dig in, so might try again later this month when I do.

"The source described leading AI labs or their allies approaching the administration every 3-5 months with an idea to ban open-source models."

The call is coming from inside the house. Anthropic and OpenAI built their AI hype machines too fast and now are desperate to prevent price competition driving the value to $0.

AI is going to be ubiquitous like the spellcheck running in my browser and at or slightly above the cost of electricity to run the models.

The blog post says they will release them July 27

Sharing some tips and tricks I found for staying in a flow state while coding with LLMs. Somtimes I found myself being exhausted from using LLMs and lately I've been trying to be more deliberate in the way that I use LLMs to help with coding.