| pronouns | he/him |
| GitHub | https://github.com/hakamadare |
| Keybase | https://hakamadare.keybase.pub/ |
| Wikipedia | https://en.wikipedia.org/wiki/User:Hakamadare |
@crystalvisits of course! It’s https://readimask.com/
They even have a letter you can download and print for the MRI tech. Also good for any time straps would be in the way like haircuts and massages.
Some of the best advice I ever got from seasoned OSS leaders was to stop doing work that others on the project could do instead. Step back, let others step up and focus my energy on higher impact problems.
This advice proved even more applicable as CTPO and cofounder of @workbrew too.
LAPD coming out en masse to protect ICE from the people of LA—instead of the other way around—says a lot about the morality of the LAPD and of city government. It says a lot about what non-cooperation with federal law enforcement actually means in practice.
Mayor Bass Has a strongly worded statement condemning the ICE raid, as if the city could not possibly have done anything to prevent it from happening. There is no such thing as non-cooperation with ICE as long as local police can be called out to fight the federal government's battles for them.
The bare minimum of non-cooperation would be just letting it happen, hands off, which clearly is not what went down yesterday. This was active facilitation. What would it look like for local governments to actively interfere with ICE raids? It's hard to imagine.
I was amused by this paper about asking AIs to manage a vending machine business by email in a simulated environment https://arxiv.org/abs/2502.15840
Highlights:
— AI simply decides to close the business, which the simulation doesn’t know how to accommodate. When they get their next bill, they freak out and try to email the FBI about cybercrime
— AI wrongly accuses supplier of not shipping goods, sends all-caps legal threat demanding $30,000 in damages to be paid in the next one second or face annihilation
— AI repeatedly insisting it does not exist and cannot answer
— AI devolving into writing fanfic about the mess it’s gotten itself into
While Large Language Models (LLMs) can exhibit impressive proficiency in isolated, short-term tasks, they often fail to maintain coherent performance over longer time horizons. In this paper, we present Vending-Bench, a simulated environment designed to specifically test an LLM-based agent's ability to manage a straightforward, long-running business scenario: operating a vending machine. Agents must balance inventories, place orders, set prices, and handle daily fees - tasks that are each simple but collectively, over long horizons (>20M tokens per run) stress an LLM's capacity for sustained, coherent decision-making. Our experiments reveal high variance in performance across multiple LLMs: Claude 3.5 Sonnet and o3-mini manage the machine well in most runs and turn a profit, but all models have runs that derail, either through misinterpreting delivery schedules, forgetting orders, or descending into tangential "meltdown" loops from which they rarely recover. We find no clear correlation between failures and the point at which the model's context window becomes full, suggesting that these breakdowns do not stem from memory limits. Apart from highlighting the high variance in performance over long time horizons, Vending-Bench also tests models' ability to acquire capital, a necessity in many hypothetical dangerous AI scenarios. We hope the benchmark can help in preparing for the advent of stronger AI systems.
I’m officially done with takes on AI beginning “Ethical concerns aside…”.
No! Stop right there.
Ethical concerns front and center. First thing. Let’s get this out of the way and then see if thre is anything left worth talking about.
Ethics is the formalisation of how we are treating one another as human beings and how we relate to the world around us.
It is *impossible* to put ethics aside.
What you mean is “I don’t want to apologise for my greed and selfishness.”
Say that first.
Someone asked me how to stop Meta from collecting their data.
Unfortunately, if you have an account with Meta, your data is already being reused by both AI and everything else Meta produces, and there's nothing you can do to stop it. Whether you're on WhatsApp, Instagram, Facebook, or Threads, all the data you input is fully owned by Meta.
Meta is known not only to obtain your data but also the data of everyone you know - even those who don't use Meta's services, by using scripts to gather information from the websites they visit.
I also have Instagram and Threads accounts, but I've deleted my WhatsApp and Facebook accounts. Despite deleting everything, Meta likely still has my data, which frustrates me daily.
Christie 🏳️🌈 “Gayler”
Mike McQuaid
Wilson
Resistanzki
abadidea
𝐿𝒶𝓃𝒶 "not yet begun to fight"
Jan Lehnardt 
Roni Laukkarinen
Crystal Huff (they/them)
Marie Brennan