How can someone say they have a disease or affliction while not understanding what exactly that affliction is as opposed to what they think it is/have been informed by media/society/non-professionals?
No, sandstorm was deprecated years ago. I was building a platform that was similar to/inspired by sandstorm, and I closely followed it until its creator deprecated it. They now work for Cloudflare and released the 'WorkOS' project a couple weeks ago through them/for them.
(Not disagreeing with your point, pointing something else out) What happens when flock sells the company + data? That collected data doesn’t get deleted and could be used for whatever purpose at that time, as the company could be dissolved along with the contracts.
That’s not how data agreements work. The city/entity would need to agree to the revised contract. Is it possible? Sure. Is it something I am concerned about? No, I would rather focus on the facts.
In California, it's thought that only 9% of data brokers are compliant with state laws requiring transparency and deletion of personal data.[0] It only goes down from there; most states do not have strong consumer privacy laws.
Furthermore, most municipal entities are not in a position to litigate a case over these changes for an extended period of time, and Flock is changing their terms to make it harder to terminate contracts. They can also now keep the data as long as they'd like. [1]
This, along with general corporate conduct these days, makes it very likely that they're actively planning to sell the data and might be doing so already in.
The article says otherwise. It opens with it isn’t genocide but is crimes against humanity.
The U.S. State Department’s Office of the Legal Advisor concluded earlier this year that China’s mass imprisonment and forced labor of ethnic Uighurs in Xinjiang amounts to crimes against humanity—but there was insufficient evidence to prove genocide, placing the United States’ top diplomatic lawyers at odds with both the Trump and Biden administrations, according to three former and current U.S. officials.
Not them but my understanding is that the harness will send a simple 'heartbeat' message to keep the cache 'warm', (see prefix caching: https://handbook.modular.com/inference-optimization/prefix-c... ) which can then be edited/changed, which does cause the user to incur a fee, but its much less than the amount they'd pay on a no-cache hit request.
I worked at Apple on Schoolwork for the last few years of my career and saw firsthand how Apple handles data it gathers. As an example, every device was given a unique identifier that in no way identified the user of the device (I mean it was simply a UUID). Additionally, it was tossed and a new one recreated every 11 months.
Because I am inclined to give Apple the benefit of the doubt (your mileage may vary), I am assuming the binary hash they send is intended to protect users from malicious binaries (once their hash is identified of course). And if Apple in this case also rotates the UUID every 11 months, I don't have to worry about them targeting me and my habits specifically (well, not beyond 11 months windows in any event).
Additionally, when at Apple, we had privacy teams walk through things like error logs that our app and framework wrote—making sure there was no PID (personally identifiable information) in the logs. A naive engineer might, in a URL-fetch timeout, for example, log something like, "Timeout for URL request: 'myblog.blogsite.com'". Privacy would ask, "Is it important we log the full URL? Might we instead just log the domain?"
I think those are both different in that they require the user to say they trust this code. Additionally the first is arguably not a bug (the code in tasks.json will indeed run if you say you trust the project) and the second was reported and fixed within two months.
Yes, my point was more so the underlying behavior of automatic execution of binaries associated with/inside of a git project exists already within vscode as a pattern, and so seeing cursor doing the same wouldn't be surprising.
Why make this comment without having tried it first? It very clearly is not useless and performs a lot better than one might expect. I am currently waiting to do more benchmarks of it in comparison to the full weight model, but it seems promising/better than Mistral Nemo at a lower file size.
I think what OP means is that the "minimum viable product" for a daily use LLM is probably somewhere around e.g. GPT 4o's level of intelligence (YMMV). Below a certain threshold, you are better off using specialized machine learning models rather than general purpose LLMs. It's very difficult to get that level of intelligence fully local on a mobile device without streaming to the cloud.
I do think this would be interesting if they made these easy to finetune, as I do think this level of intelligence is likely sufficient for many applications and could be extremely cheap to run.
reply