After they were roasted by the 2022–2023 Pentagon doc leaks, it was pretty obvious they were going to take action.
And not just that event: Parents are roasting Roblox for kids getting groomed, but after the relationship is initiated, the groomers always immediately the convo to Discord.
Did he argue SF likes graffiti? I don't think he does, and the people living in the city certainly don't. These are criminals tagging buildings, and city officials who either don't care or are too busy with other things. I'm not aware of anyone who actually lives there who likes graffiti, and logically there's no reason anyone should. If someone wanted a mural they would have hired a real artist to do it.
He's arguing that the authorities aren't doing anything about it, and the reason is, (going out on a limb here) SF residents are sympathetic to the renegade artistic expression argument.
> SF residents are sympathetic to the renegade artistic expression argument.
SF residents are incredibly snobby when it comes to street art. The typical tagging, 2 minute stencil sprays, and so forth are not up to posh standards of SF residents. I don't think most SFers think those are "renegade artistic expression". Maybe some of folks in Berkeley would but not SF.
There's a huge disconnect from the city residents and a lot of what happens by the government. SFPD is a prime example of this. Almost none of the cops live in SF. A lot of the people committing crime also don't live in SF. It's a weird city.
I've lived through this whole epidemic, I remember seeing the origional modern Grafitti guy's work all over North Philly on my way to the doctor in the mid 1970s. That guy's name was CornBread. The artful mural like tagging origionating in the Bronx is another story but there have always been a-holes who tag some scribble over other stuff.
If there were community areas that were designed for painting, that would be totally fine by me. A big wall that is painted white, maybe with some ladders nearby if that doesn't violate health and safety rules, and tell people to go nuts. Though you would potentially get a lot of disagreeable content, but I suspect that they would quickly get overwritten anyway
There are many such walls or discreet hidden spots, like tunnels, where the graffiti artists paint for other artists. But painting in wide-open spots where the public can judge is just part of the culture.
Maybe I should have compared graffiti to the release of Gnutella.
Follow up Q: what are you supposed to do when the context becomes too large? Start a new conversation/context window and let Claude start from scratch?
Context filling up is sort of the Achilles heel of CLI agents. The main remedy is to have it output some type of handoff document and then run /compact which leaves you with a summary of the latest task. It sort of works but by definition it loses information, and you often find yourself having to re-explain or re-generate details to continue the work.
I made a tool[1] that lets you just start a new session and injects the original session file path, so you can extract any arbitrary details of prior work from there using sub-agents.
Either have Claude /compact or have it output things to a file it can read in on the next session. That file would be a summary of progress for work on a spec or something similar. Also good to prime it again with the Readme or any other higher level context
It’s a good idea to have Claude write down the execution plan (including todos). Or you can use something like Linear / GH Issues to track the big items. Then small/tactical todos are what you track in session todos.
This approach means you can just kill the session and restart if you hit limits.
(If you hit context limits you probably also want to look into sub-agents to help prevent context bloat. For example any time you are running and debugging unit tests, it’s usually best to start with a subagent to handle the easy errors. )
It feels like one could produce a digest of the context that works very similarly but fits in the available context window - not just by getting the LLM to use succinct language, but also mathematically; like reducing a sparse matrix.
There might be an input that would produce that sort of effect, perhaps it looks like nonsense (like reading zipped data) but when the LLM attempts to do interactive in it the outcome is close to consuming the context?
I ask it to write a markdown file describing how it should go about performing the task. Then have it read the file next time. Works well for things like creating tests for controller methods where there is a procedure it should follow that was probably developed over a session with several prompts and feedback on its output.
Start in plan mode, generating a markdown file with the plan, keep it up to date as it is executed, and after each iteration commit, clear the context and tell it to read the plan and execute the next step.
Yes you can literally just ask Claude Code to create a status line showing context usage. I had it make this colored progress bar of context usage, changing thru green, yellow, orange, red as context fills up. Instructions to install:
Here's an example. One of my friends works for a manufacturing company. He attended a protest. The next day ICE called his employer and he was informed that if he attended another protest he would be fired. All this b/c he had a small company logo on his jacket.
The ability to en-mass record, lookup and intimidate citizens is unprecedented and while I have no hard proof that this is due to Palantir, it sure smells like it
ICE employees working to “injure, oppress, threaten, or intimidate any person in any State, Territory, Commonwealth, Possession, or District in the free exercise or enjoyment of any right or privilege secured to him by the Constitution or laws of the United States, or because of his having so exercised the same”, such as exercise of 1st Amendment rights, is itself super-illegal (not just “outside their authority” but a federal criminal violation of the KKK Act), whether or not the employer’s response to the resulting pressure is also illegal.
My understanding (and I couldn't get past the app paywall) is that Palantir is joining databases from many different federal and state agencies, including passport and driver license photos. The app then allows you to scan a phase and it finds a match. It returns information on the person found, including citizenship.
The existence of this technology means that ICE can grab anyone they want, scan their face, and instantly have (or not have) probable cause to arrest them. Without the app, there would be hours before probably cause could be established which makes justifying the detainment legally much harder. I.e without the app, ICE has to actually build a case or see something suspicious for each target. With the app, ICE can just mass sweep people.
Which should be illegal, but thanks to the shadow docket order on Noem v. Vasquez Perdomo, is happening anyway.
CBP has been taking photos of all legitimate foreign visitors to the US for over 20 years. I presume any catch-and-release border apprehensions are subject to the same photographs.
How hard is it to do facial recognition on just this dataset in real-time?
Eh, joining these datasets can be challenging. Names can be spelled differently or changed, dates of birth can be off, people can share names and dates of births, addresses change and are can be expressed in multiple ways, databases may store names as a single string or separate fields, middle names may be missing or initials, databases might not share IDs etc. So it's kinda hard to do well although nothing really exciting technology wise.
This, incidentally, is why the "confidence score" is needed. And why the app frequently gets data (including citizenship) wrong.
Overeating doesn’t just happen in a bubble - there is a confluence of issues creating anxiety and stress in Americans’s daily lives leading to the obesity issue.
I agree with this take too. Culture influences it but I don’t think anyone can be singled out being immune. We are stressed and overworked. Simple food takes work. We no longer have a village. If you have kids so many are just stuck at home.
I compare it to my childhood which was a while ago but not that far and I would go out in the middle of the day with instructions to come back home before dark. I would be running all over town on my bicycle. Now parents in the US are obsessing over travel sports and keeping booked calendars for their kids. Both parents will be working. There is nobody around put a meal together.
And not just that event: Parents are roasting Roblox for kids getting groomed, but after the relationship is initiated, the groomers always immediately the convo to Discord.