One simple way is it has superseded google/SO on my priority list for some reference lookups on topics I need to dabble in for work. (Like “Django custom function generate uuid”) Practically speaking, I’ve found this (most of the time) reduces the cognitive overhead of context switching. Less clicking, sifting, looking for recent answers, etc.
For slightly more complex reference on topics I’m dabbling in (Like “what’s the right way to use React hooks for this component”) the benefit of leveraging ChatGPT increases. In many cases you’re able to skip the step of generalizing your problem and you can address your needs from the “inside out”.
Inside out example: starting at some API method names and sample code for some libraries, rather than starting with a list of libraries or worse not knowing what kind of library you need yet.
Finally the place where it has had the most impact for me is as a force multiplier on efforts outside of my comfort zone as an engineer. If I have a side project idea, or if I’m trying to be ambitious in design at work, I can often discuss approaches at length with ChatGPT and walk away with several ideas, or even conclusions about how best to proceed without exhausting myself.
Tangential but may be of interest: a cool book (compilation?) that somewhat addresses this topic is ‘Ideas That Created the Future’ edited by Harry R. Lewis [0].
Working through it now - reading the Laws of Thought by George Boole was a spiritual experience for me. The connection between natural language, logic, and math finally made sense to me. Before I had a hard time understanding the saliency of things like Analytic Philosophy (what does language have to do with truth?) or why the work of someone like Chomsky would be so important to computer scientists. Boole wasn't the only person at time to work on a project to reduce logic to solvable equations (clearly that time in history was ripe for this work, and also Leibniz had laid some of the spiritual groundwork, himself inspired by Ramon Llull and Confucian texts) but the way he wrote about it really resonated with me.
The other surprising thing to me was that it didn't really resemble modern algebra at all. As a former math student I think it's easy to create a mental model of math history where when new fields are created they resemble the modern incarnation. But when you actually look at the primary founding texts its clear the ideas and notation are pretty unrefined at least compared to today's standards.
Is this a summary or the entire text? I think I saw messaging from Coinbase in anticipation of this proposal earlier this week. [1]
Not a lawyer but the definitions and language seem a little shallow to me considering how lofty the proposal of a new regulatory agency is.
Wouldn’t one expect a higher level of detail, supporting evidence, citations or legal precedent from consultations with “30+ crypto firms, 25+ members of congress and/or staff, 4 major law firms, and 3 trade groups”?
What is the worst case scenario that the author imagines happening to Rust due to Amazon’s influence (or anyone’s influence)?
This is unfamiliar to me - are there notable stories of what poor steering has done to a programming language? What were the consequences?
I can parse the basic grievances here on power dynamics, losing control of something you helped build…that’s familiar enough…but I’m having trouble understanding the gravity of what the author is passionate about preventing.
(In case it isn’t obvious I’ve never been a long term contributor to an open source project - genuinely curious about the context here)
This looks cool! I spend half the day in terraform and AWS and I’m looking forward to trying this out later today.
Regardless of how that goes, it’s really nice to see a project that addresses this specific area. Way to go!
The AWS console is (at least for the popular services and API endpoints) super effective.
That said, when you’re deploying a project that has components that span almost a dozen services in AWS that can quickly become a dozen tabs just to assess the state of the project. Not very fun - looking forward to seeing how this handles that.
So there actually is a very quiet functionality hiding in MacOS Notes.app I noticed recently.
If you take any screenshot to clipboard that includes some text, then paste it into Notes, it will silently name the resulting image file using the text pictured.
Not terribly useful but I did find it helpful once while taking screenshots for documentation.
In the past I’ve felt the opposite of your sentiment.
I’ve found my email to be unwieldy. I’ve felt like I have to apply constant hygiene and maintenance to keep it from becoming useless.
Checking out on an e-commerce site, trying out a new tool, registering a free trial...these are things I guess I do frequently enough that it seems to doom me to sort and sift through noise.
I know the common solutions are a good set of filtering rules or separate emails/aliases. I lean toward the latter and it’a definitely helped but I still find it all a bit of a pain and time sink.
That said, you make a nice case for it and I can definitely identify with some of the ‘zen’ of email you’re describing.
I feel like I’ve tried every email client out there, but is there some solution I haven’t thought of of where I get a friendly GUI but can configure mail filtering in code (i.e something I can keep in git) rather than clicking around?
I’ve toyed with ideas of bundling by sender - or quietly placing new senders into their own ‘message requests’ type bin for me to approve or deny.
Or some kind of intelligence for identifying, for example, a receipt or confirmation email (which I want indexed) vs a promotional email from the same vendor (which I do not).
> I feel like I’ve tried every email client out there, but is there some solution I haven’t thought of of where I get a friendly GUI but can configure mail filtering in code (i.e something I can keep in git) rather than clicking around?
A couple jobs ago, before I gave up on processing all my mail, I had a perl script that grabbed headers from IMAP (with caching, because Exchange is slooooow), and filtered with code (basically a list of pairs of code refs that took a message => what folder to move the message if the code returned something truthy (or maybe the code returned a folder... not sure anymore). This let me do things that are hard in server side filters like moving things only when it's read.
Not opting into marketing spam (or opting out immediately in the website's settings) takes care of 99% of it for me. On a set of ~250 accounts I will maybe receive one marketing e-mail per month (and they quickly get a complaint to their data protection officer to discourage that happening again).
For transactional stuff a set of rules to automatically read+archive stuff that you need to keep (invoices, receipts, order notifications, etc) but don't necessary need to action.
This leaves you with only the relevant stuff in the inbox and makes it much easier to do "inbox zero". I will typically get less than one email a day actually landing in my inbox using this technique.
> Or some kind of intelligence for identifying, for example, a receipt or confirmation email (which I want indexed) vs a promotional email from the same vendor (which I do not).
Yeah, one thing where even statistical filtering sucks. Usually there's not enough confirmation vs promo samples to train a filter and some senders tend to use the same templates for both.
1. Mark them all as spam. I never accept newsletters so I will consider every newsletter email as spam. They don't last much in my inbox.
2. Try inbox zero and just archive emails aggressively directly from your phone’s notification without even opening email.
My inbox currently has 10 emails that one day I'll get to and the rest of spam and transactional email is just gone from view within the hour or day.
I wish I could have an "important" and a "transactional" email view, but that's not going to happen without a lot of work, so email search still sucks.
> it doesn't make sense to say email search sucks.
Yes it does, because in almost every case it sucks :) The only one where it doesn't is where it's all indexed by Google (the bad option) or where you index it all locally (e.g. with mu). The second is a slightly less bad option but not by much.
> Yes it does, because in almost every case it sucks :)
I mean I get it but fundamentally it's not true. Might seem like a nitpick but it's not. If slack search sucks, that's an absolute statement. It is what it is and there's nothing you can do about other than beg a product manager at slack to make it better but they'll likely ignore you.
With email, an open standard, you can just switch clients. Or bypass clients entirely and handle search separately. There are no limits to what you can do.
It sucks because transactional emails still show up next to important emails. What’s a client that separates this without me manually doing so (via filter or whatever)?
Gmail is good at this when splitting the inbox, but after that it’s all forgotten (unless, once again, you add a manual filter to exclude such emails)
The idea of Blackberry's unified list really resonates with me! And I strongly endorse your two tips, they really help.
About your important/transactional split: shameless self plug but I'm working on a project that's all about a unified list, and one part of that is splitting incoming into important and not important. If you check it out, feedback welcome! garrett@[productname].com
For slightly more complex reference on topics I’m dabbling in (Like “what’s the right way to use React hooks for this component”) the benefit of leveraging ChatGPT increases. In many cases you’re able to skip the step of generalizing your problem and you can address your needs from the “inside out”.
Inside out example: starting at some API method names and sample code for some libraries, rather than starting with a list of libraries or worse not knowing what kind of library you need yet.
Finally the place where it has had the most impact for me is as a force multiplier on efforts outside of my comfort zone as an engineer. If I have a side project idea, or if I’m trying to be ambitious in design at work, I can often discuss approaches at length with ChatGPT and walk away with several ideas, or even conclusions about how best to proceed without exhausting myself.