chatmasta

7 ms·

61.
by chatmasta 4mo ago
But it’s not that different from the whole premise of their red team scaremongering which was “we pointed the model at a source file and told it to find an exploit.”
62.
by chatmasta 4mo ago
Anti-jailbreaking and passport verified access to model families.
63.
by chatmasta 4mo ago
It’s a pretty safe bet that every frontier lab has multiple foreign intelligence agencies running assets inside of it.
64.
by chatmasta 4mo ago
So are guns, which we constitutionally protected. In fact there’s probably a decent argument that AI should fall under 2nd amendment protection.
65.
by chatmasta 4mo ago
It’ll be sorted out after OpenAI releases their next model.
66.
by chatmasta 4mo ago
Even Amazon does this. So scummy. This exchange rate trick should be illegal everywhere.
67.
by chatmasta 4mo ago
It’s logical consideration with “logical” meaning Spock style logic, ie utilitarianism at all costs. Another prominent EA is SBF for example. It’s designed to sound innocuous and many of its cultish promoters may genuinely believe it’s inno
68.
by chatmasta 4mo ago
The company was founded basically out of the effective altruism movement.
69.
by chatmasta 4mo ago
Doesn’t PgDog also handle the sharding by proxying the writes? Maybe I missed something but I thought this is their value prop. It’s not just another PgBouncer.
70.
by chatmasta 4mo ago
The composition of the average transaction will be different in a shopping cart (lots of writes and updates) compared to your use case which sounds like it skewed read heavy. With Postgres it’s generally easier to scale reads because it doe
71.
by chatmasta 4mo ago
What if the data goes to Anthropic EU servers?
72.
by chatmasta 4mo ago
Yeah it’s wild. But I think the attitude is more like “everyone is taking the same risks together, so we won’t be alone if the ship hits an iceberg.” Nobody got fired for buying Frontier AI.
73.
by chatmasta 4mo ago
Anthropic has all the answers for that. You’ll go through some compliance exercises and classify them as a subprocessor of highest tier of data sensitivity.
74.
by chatmasta 4mo ago
They give it some thought, but Anthropic and AWS have the whole menu of compliance and security checkboxes needed to reassure CISO it doesn’t need to be “the office of no” and can allow the AI onboarding. The pressure to adopt and adapt to
75.
by chatmasta 4mo ago
Follow the Yugoslavia model and ship soldiers from each country to live and work with each other in their first years in military service. Then again… Yugoslavia maybe not the best example here…
76.
by chatmasta 4mo ago
I see some deeply enthusiastic comments like this whenever it’s mentioned. I wonder how much lack of adoption is simply lack of awareness. Apple needs to find ways to make experimenting with it, and experiencing specific use cases, more app
77.
by chatmasta 4mo ago
They will, but they won’t realize it’s writing code. It will look like Claude Cowork, which writes code for itself under the hood but is results-oriented for the user.
78.
by chatmasta 4mo ago
What? They make money from their own inference and models too, which they can train effectively for free by funding their operations with rental income from their last gen datacenter.
79.
by chatmasta 4mo ago
xAI covers their cost of N-1 datacenter while running their own models in N and building out N+1.
80.
by chatmasta 4mo ago
I’d like to see less focus on Playwright and more focus on giving the agent more than just an MCP to browser automation. Make it multi-modal, figure out how to optimize when to send screenshots to which model, etc… current coding harnesses
81.
by chatmasta 4mo ago
Luckily for Apple, their primary competitor has managed to make their OS even worse. Windows becomes more unusable with every update, and on top of that, continues shoving telemetry and advertising into the OS. I installed windows on a lapt
82.
by chatmasta 4mo ago
Try messing with the settings more. Grayscale (or whatever it’s called) makes the transparency much more tolerable, or even _nice_, than with multi-color. It’s also more palatable on iOS because you only have one window open at a time. Many
83.
by chatmasta 4mo ago
I’m not sure they need novelty or moat. AI compute resources are so scarce that inference providers will buy whatever is available. SpaceX sells inference hardware in bulk, with a proven track record of running inference and training worklo
84.
by chatmasta 4mo ago
Grok does seem to have the best searching capabilities, and not just for twitter. I wonder what search engine they’re using on the backend.
85.
by chatmasta 4mo ago
Datacenter operators who rent space are selling electricity. SpaceX is selling a fully built datacenter with compute designed for a specific purpose. They’re operating at a higher level of the value chain and can charge accordingly.
86.
by chatmasta 4mo ago
Also weird how it’s all or nothing. Feels like it should just be a theme you choose for your OS.
87.
by chatmasta 4mo ago
When I download a new app on iOS, I immediately disable searching it and disable background updates. I hate the search feature and only use it as an app launcher when I’m accidentally not on the App Library screen. On macOS, I also disable
88.
by chatmasta 4mo ago
Liquid Glass is fine for me since I put it in grayscale. I actually like it. I’m just talking about iOS though. Haven’t updated to Liquid (Gl)ass on macOS yet.
89.
by chatmasta 4mo ago
An open mind to not having a knee jerk reaction to “this is vibe coded therefore it’s not production-ready.” The delta between the prototype and production is likely far lower than you think, and many engineers are sneering at the prototype
90.
by chatmasta 4mo ago
> business people don't understand why you can't just put their app in production I’d flip that around and say that engineers don’t understand that sometimes you _can_ just put their app into production. It might take some clea

More ›