Are the documents encrypted? Is privacy a focus? One major reason that people want to move away from Google Docs and Notion is because of their very invasive privacy policies.
They are not and while basic privacy and security is definitely a focus (making sure random external people can't get access to your docs and account) end-to-end encryption and privacy is not. There are tools that focus on that that seem great. I think proton mail has a product in that space, and https://docs.fileverse.io/ seems interesting, and there are many more on the more notion side. We wanted to focus our efforts on the interface and experience.
All that being said, I'd love at some point to manage to get everything end-to-end encrypted, but it'll be a big lift!
If I understand correctly, you’re doing the sync client-side, meaning you don’t need the payload in cleartext server-side. In that case, it should be relatively straightforward to generate an encryption/decryption key and put it into the fragment part of the URL, so that the server doesn’t see it, and encrypt all payloads with it.
There's actually a fair amount going on server side. For one, we have server authoritative sync. There's a lot of mutations that would be tricky to reason about in purely p2p sync, and just applying them to an authoritative source makes that a lot simpler (and is what let's us use Replicache!). Beyond that though we have a bunch of features, custom domains, polls, rsvp blocks, that depend on us being able to control permissions.
It's definitely possible to do all if this in a p2p context, but it's much harder and increases the complexity a ton. It also makes it harder to be fast, you're naturally increasing the amount that needs to be done client side, and server rendering is out the door. There's a bunch of client-side editors that I think are great, and end-to-end encryption is definitely something I want to work towards, but it's out of scope for us right now.
I’m watching the horror show of US politics unfold from Western Europe, where I know every outrageous step that Trump takes, unchallenged, will only empower and encourage those would-be dictators who are attempting to take over and possibly dismantle our political systems too. But to call this “important news” isn’t really a fair description. It’s an opinion piece, which at surface level (hopefully) echoes the opinions of many readers who are recoiling in horror about the apparent destruction of democracy before their very eyes; however it is also a propaganda piece from one extremist celebrity leader against another, probably just as selective in its facts as all the guff that comes from the “other side”. The solution to right-extremism is not left-extremism, the real question is how can we talk reasonable people at both ends of the spectrum back from the ledge and attempt to form a new era of collaborative politics that aims for balance and fairness?
The biggest problem is that we'd like our leaders to be exceptionally smart in order to efficiently govern the country, but at the same time we don't want the leaders to outsmart the voters. These two expectations are mutually exclusive, and most democracies just end up praying that whatever politicians get elected, they turn out to be benevolent, instead of focuing on self-interest. Whoever wields the power, how do we keep them in check, and what kind of power do we use for that?
In the US they have the check & balances mechanism to ensure against "Who will watch the watchmen" problem. Still with the current president they are already being put through their paces. The constitution could be weakened by amendments or even disregarded if you have enough power in your grip.
Musk was on stage with Trump and said he wanted to slash spending and the US people voted them in winning every swing state. Not sure it's a good idea but that's what 'we the people' chose.
If the democrats earnestly cared about USAID they wouldn’t have painted a giant target on it by treating it as a slush fund for partisan pet projects.
Trump and Musk were crystal clear ahead of the election that they intended to stop exactly this type of partisan funding, and the public agreed at the election.
While I can appreciate there are some legal ambiguities around the specific techniques used to enact that mandate, the left appear to be attacking the mandate itself, which is undemocratic and risks creating the impression that they are attending to subvert the will of the public.
Watch this [0] interview with a former USAID diector. He states you dont blow up the organization and attack the non-political civil servants, you install a director that works with you on your policies. Also their IG was investigating starlink, which may be related and is certainly a conflict of interest [1].
The Starlink contract is almost certainly immaterial. It's only a temporary arrangement to support Ukraine and the contract value is under $1 million per year.
If you remember Starlink even initially provided the hardware and services to Ukraine for free.
nLine is a startup improving electricity reliability in developing countries through innovative sensor data and analytics. Our small team (10+) is building robust cloud and data systems that drive better decision-making for power grid improvement.
Role: Platform Engineer – Cloud & Data Systems
Must-have:
- 3+ years’ experience engineering with back-end focus and Python skills
- Deep experience with GCP, Terraform/Kubernetes, and distributed computing
- Proficiency with PostgreSQL/TimescaleDB and big data processing
Nice-to-have:
- Experience with Databricks, Delta Lake, and Helm
- Some background in IoT or power grid operations
You will design and optimize data pipelines, manage cloud resources, and support critical infrastructure tasks. We're looking for an independent innovator who complements our global team's skillset.
What is the cost of planning and working through, let's say, a manageable issue in a repo? Does it make sense to use 3.5/Sonnet or some lower cost endpoint for these tasks?
It's hard to put a precise number on it because it depends on exactly how much context is loaded, how many model responses the task needs to finish, and how much iteration you need to do in order to get the results you're looking for.
That said, you can do quite a meaty task for well under $1. If you're using it heavily it can start to add up over time, so you'd just need to weigh that cost against how you value your time I suppose. In the future I do hope to incorporate fine tuned models that should bring the cost down, as well as other model options like I mentioned in the post.
You can try different models and model settings with `plandex set-model` and see how you go. But in my experience gpt-4 is really the minimum bar for getting usable results.
One issue I've struggled with in Zed is full language support (linters/formatters). Has anyone at the Zed team thought about integrating a metalinter like Trunk Check (https://docs.trunk.io/check)?
