you probably say this jokingly, but this is not a bad take. or rather, "assume every system on earth can/will be breached". that is a good motivation to improve the security for your system, if you have the resources to.
I'm using an M3 Ultra w/ 512GB of RAM, using LMStudio and mostly mlx models. It runs massive models with reasonable tokens per second, though prompt processing can be slow. It handles long conversations fine so long as the KV cache hits. It's usable with opencode and crush, though my main motivation for getting it was specifically to be able to process personal data (e.g. emails) privately, and to experiment freely with abliterated models for security research. Also, I appreciate being able to run it off solar power.
I'm still trying to figure out a good solution for fast external storage, I only went for 1TB internal which doesn't go very far with models that have hundreds of billions of parameters.
In theory, born into it. That was just a foil to put an air of legitimacy over the institution.
In the real world, there was (and is!) an incredible power game over who decides over what, who gets to live, who must abdicate, how much the real power lies with the King and how much with aristocracy or the Church and so on. It's a constant rebalancing of power factors.
I mean, it depends what you mean by "engagements." He was invited to meet Epstein via an intermediary -- but he blew off the invitation and never had any contact with Epstein.
In general, Epstein was fond of "collecting" scientists who might entertain his clientele and house guests at parties.
>The fix was to use GCC as an online known-good compiler oracle to compare against.
>This was a clean-room implementation (Claude did not have internet access at any point during its development); it depends only on the Rust standard library.
How does one re-conciliate both of this statements? Sure one can fetch all of gnu.org in local, and a model which already scrapped the whole internet somehow already integrated it in its weights, didn’t it?
The worldwide median household income (as of 2013 data from Gallup) was approximately $9,733 per year (in PPP, current international dollars).
This means that $20,000 per year is more than double the global median income.
A median Luxembourg citizen earns $20,000 in about 5 to 6 months of work, a Burundi one would on median need 42.5 months, that is 3.5 years.
OAI is running boundary pushing large models. I don’t think those “second tier” applications can even get the GPUs with the HBM required at any reasonable scale for customer use.
Not to mention training costs of foundation models
Spanish don't mess around in this matter. You have outright Prosegur or some other security company's cameras openly in your Booking/Airbnb apartment, and a word of the host that "it's not recording".
Depends on what we categorize as a coding agent. Devin was released two years ago. Cursor was about the same, and it released agent mode around 1.5 years ago. Aider has been around even longer than that I think.
For the same reason they use "tokens" instead of kilobytes: so that you don't do the conversion yourself and realise that for example spending a million "tokens" on claude-opus-4.6 costs you anywhere from $10 (input tokens) to $37.5 (output tokens). Now, 1 million tokens sounds pretty big until you realise that's about 4 megabytes of text. It's less than three floppy disks of data going back and forth.
Now let's assume you want to send a CD worth of data to Opus 4.6. 700 megabytes * $10 (price per million input tokens) / 4 (rounding down one megabyte to roughly 250k "tokens") = $1750. For Opus 4.6 to return a CD amount of data back to you: $37.50 * 700 / 4 = ~$6.5k.
A terabyte worth of data with a 50:50 input/output ratio would cost you $5.7 million. A terabyte worth of data with a 50:50 input/output ratio on gpt-5.2-pro would cost you $25.2 million. (Note: OpenAI's API pricing still hasn't been updated to reflect 5.3 prices.)
So we get layers upon layers upon layers upon layers upon layers of obfuscation to hide those numbers from you.
Related to "Why So Many Control Rooms Were Seafoam Green" - Soviet designers apparently reached the same conclusion, but they applied it to aircraft cockpits instead of control rooms and used a slightly more blueish color: https://aviation.stackexchange.com/questions/16434/why-are-r...
not quite as technically rich as i came to expect from previous posts from op, but very insightful regardless.
not ashamed to say that i am between steps 2 and 3 in my personal workflow.
>Adopting a tool feels like work, and I do not want to put in the effort
all the different approaches floating online feel ephemeral to me. this, just like for different tools for the op, seem like a chore to adopt. i like the fomo mongering from the community does not help here, but in the end it is a matter of personal discovery to stick with what works for you.
The British Crown had to concede some rights centuries ago, or there would have been civil wars and probably no more crown. Dear Leaders are the ones that don't have to concede anything, yet.
It's a tricky issue. In many countries it's not illegal and quite common for children to run around naked in public, during the summer on beaches for example, and so millions of people have holiday photos that are technically CSAM in their possession that they don't even know they have.
I don't think this will be an issue, given history. COBOL was developed so that someone higher up could use more human language to write software. (BASIC too? I don't know, I wasn't around for either).
More modern-day, low/no-code platforms are advertised as such... and yet, they don't replace software developers. (in fact, some projects my employer does is migrating away from low/no-code platforms in favor of code, because performance and other nonfunctionals are hidden away. We had a major outage as a result when traffic increased.)
This is an interesting word. Halfway across the Alpo-Himalayan backbone the word for this tool is "kobilitsa", cognate to "kobila" (female horse) and I always figured it was a metaphor for the arched form of the bar (and probably referring to how women were stuck with the task of fetching water with it).
But now it seems the word "kaavidi" has reached our ancestral lifestyle all the way from the Indian lands! And the transformation it underwent was more due to how people "normalize" foreign words; same linguistic churn that gives us backronyms and false etymologies (see also "eggcorn"). Whoa
I presume linguists have already studied the naming of household implements when deriving Proto-Indo-European. I've never encountered much literature on the subject, nonetheless I find the subject rather fascinating.
> I avoid software that I know to be problematic. Also, I firewall off all outgoing network traffic by default, and whitelist very sparingly.
how do you stay informed about what software is problematic? what are examples of addresses you whitelist? if you feel comfortable sharing.
> For my smartphone, I run a bare minimum of apps and refuse to install new ones without an extremely good reason. I also pipe all smartphone data through a VPN I run at home, specifically so that I can run it through my firewall and make the block-by-default policy I mentioned above cover the phone as well.
what would be an example of "an extremely good reason"? what VPN solution do you have?
This is exactly what set me off in trying to figure out the visibility gap.
What’s strange is that we’re moving into a world where recommendations matter more than a click, but attribution still assumes a traditional search funnel. By the time someone lands on your site, the most important decision may have already happened upstream and you have no idea.
The UTM case you mentioned is a good example: it only captures direct "AI to site" clicks, but misses scenarios where AI influences the decision indirectly (brand mention to later search to visit). From the site’s perspective tho... yeah it looks indistinguishable from organic search. It makes me wonder whether we’ll need a completely new mental model for attribution here. Perhaps less about “what query drove this visit” and more about “where did trust originate.”
Not sure what the right solution is yet, but it feels like we’re flying blind during a pretty major shift in how people discover things.
> Don't do anything secret on the computer
i think that might be unavoidable.
> Assume every system on Earth is breached
you probably say this jokingly, but this is not a bad take. or rather, "assume every system on earth can/will be breached". that is a good motivation to improve the security for your system, if you have the resources to.
:o)