regulation_d

I follow a few landscape photographers on YouTube and really enjoy watching them go out into the field (whether it's an actual field or mountain or beach). They've got their art (at least their own style) so dialed in, that the technical details, the subject, the composition, those are almost a forgone conclusion. What matters at that point is light. And light is luck. The difference between an average image and an outstanding one lies in something beyond their control.

And when fortune smiles on them, they get positively giddy.

I really appreciate people who understand that they have to meet luck half way. Even though they've spent hours upon hours upon hours honing their craft, the thing that puts them over the hump is both unpredictable and uncontrollable.

Yeah, from my experience the more helpful conclusion is "TTS is not commoditized yet". At some point in the next 5 years, convincing TTS will be table stakes. But for now, paying for TTS gets you better results.

Copyright law: A reason that copyright trolls are less common than patent trolls is that under copyright law, works created independently are not infringing. In court, you might have to prove that you did actually create the thing independently, but I think most juries would be sympathetic to this case. "Oh, you think that the defendant combed through your giant library of millions of logos to find this one specific, rather simple looking specimen."

Also, a lot of logos are simply not "artistic" enough to be eligible for copyright. So in general, logos are more likely to be the subject of trademark litigation than copyright litigation.

Trademark law: In order to claim a trademark you must have used the mark in commerce. So a catalogue of logos not used in commerce is of no real value from a trademark perspective.

Re: 5. building things: The only thing that keeps me sane since moving into a leadership role is that the what I love about building, namely watching things that didn't exist-- begin to exist, still happens. I guess I'm more of a farmer now than a carpenter, but I'm still seeing my environment change in a way that I contributed to and take pride in. That includes not only watching our product evolve, but also watching our team grow.

We use Elixir for our primary application, with a fair amount of Python code to manage our ML pipelines. But we also need real-time inference and it's really convenient/performant to be able to just do that in-app. So I, for one, am very grateful for the work that's been done provide the level of tooling in Elixir. It has worked quite well for us.

i agree that with statements aren't the easiest to understand, especially for a beginner, but I think the value in having the entirety of the happy path in the initial block is very helpful for understanding the flow of the feature, once you grok the syntax.

In the first paragraph, Fred mentions and links to Erlang in Anger.

When I was coming up to speed on the BEAM this was such an amazing resource. The fact that it's free is bananas. In addition to the load management stuff, he also talked about some observability details that are really helpful (like how to more accurately understand CPU load on the BEAM). Highly recommend.

Jose, if you come across this, I just want to express my gratitude to you for Elixir and for everything you've done for the Elixir community. The thoughtfulness and adeptness you bring to the language and the community is so very much appreciated. This change is such an excellent example of the steadiness and diligence you have led with. So rarely does a single individual that I am not interpersonally close to have such an outsized impact on my day-to-day happiness. Thank you, Jose.

When I first started writing Elixir, I had no idea how much I would come to love immutable data as a language feature. "functional" can mean so many things, but for me immutable data as part of the runtime (i.e. not bolted on later) is one of the most important things. Never having to ask: did I pass by reference or value? Never having worry about data changing out from underneath you. It's pretty amazing.

Sure there are times that a need for performance calls for mutable data. But for me those times are pretty rare, and when they happen it's usually easy enough to quarantine the mutation.

There is active work being done currently to add set-theoretic types to the core language. https://www.youtube.com/watch?v=gJJH7a2J9O8

"Poor imitations" is probably harsh, but I agree with the idea that LiveView's architecture is substantially different from Hotwire or LiveWire. To have experienced Hotwire is not to have experienced LiveView, ironically enough, for reasons outlined in the posted article.

Same. At first, I thought PragmataPro was a little narrow, but now that I grown accustomed to it, everything else feels squatty or improperly kerned.

The counterpoint here is that there is already a cause of action for this type of incompetence and it's called malpractice, which is a pretty reasonable road to remedy. I don't know if you actually think these were "fraudulent court documents", but "fraudulent" actually means something very specific and this ain't it. Even if the court is considering sanctions (which is not the same as disbarment), that seems at least partially related to the attys' failure to address their failure once they were aware of it.

Something interesting about the legal profession is that it is self-regulating. The state bars are typically not government organizations. Attorneys know that confidence in their profession is extremely important and they strike the balance between preserving that confidence and, you know, destroying someone's livelihood because they don't understand how LLMs work.

Lan Lam is great! Be sure to check out her out content as well!

I mean if the OED allows for a metaphorical or hyperbolical meaning, I think you're fighting a literal uphill battle.

I'll caveat that this is mostly tailor toward SaaS companies, but it's all I really know:

I think my proxy for work-life balance is days of PTO. I'm typically going to take ~20 a year. That's what I need to recharge and avoid burnout. In a interview with a manager I would ask:

"So, in the job description it says you have Flexible PTO, what does that actually look like here?"

Usually they'll give me some sort of range or average.

A way to cut down on the chances of working for shitty leadership is to ensure you work with high performers. (not foolproof, but helpful) High performers don't like working for shitty leadership. I borrow my proxy for high performance here from the excellent book, Accelerate (Forsgren, Humble, and Kim), and look for the engineering org to implement CI/CD as well as IAC (Infrastructure As Code). I'll ask about these during question time of the technical interview, as well as other things I care about (testing philosophy, etc). I'm sure there are really great organizations that don't do CI/CD or IAC, but I would have to know _for sure_ that they were great, to work for them.

These are obviously super high level proxies for organizational health from the perspective of an engineer. But reading up on organizational health more broadly can help you identify some red flags these proxies don't cover. Check out The Five Dysfunctions of a Team by Patrick Lencioni. It's a fantastic book on organizational health.

Lastly, networking can be very challenging but also very helpful here. It's not something that comes easy to me, but it has helped tremendously in figuring out which companies in my community I do and do not want to work for. This could be less relevant now that almost everyone is cool with remote work. But even my most recent job post-covid, was something I found because of my network. Go to meetups. Go to conferences. Get coffee with people who do similar things. It can really pay off.

I've been in a similar situation. My advice is to make sure you budget in recovery time. I've had burnout that took 2 weeks to recover from, and I've have burnout that took 2 months to recover from. Just make sure you don't take on more "burnout debt" than you'll be able to repay before having to start your next thing.

I'm sure most of us have worked for shit companies that have taken advantage of us. In those cases, I think your points can be helpful.

But I'd also suggest that it's possible to find companies that align with your values, both around work/life balance and around the way the work gets done. That's certainly a skill that I've had to work to develop. And I've had some major misses along the way. But it's probably also the most important skill I've developed.

Owning your company's problems, at least a subset of them, is that fastest (only?) way to move up in a company. If you want to climb the ladder at all, you'll have to own some problems. The key is to do that for a company that recognizes and rewards that ownership. And as soon as they don't, then you become the merc and find a new thing.

After ingesting the entire tax code…

According to a quick google search, the entirety of the US tax code is over 1M words. I wonder which number GPT will support a prompt that large.

Depends what you mean by "regular cloud service", but broadly Fly.io and Heroku are examples of Platform-As-A-Service, versus AWS, GCP, Azure, and Digital Ocean which are more Infrastructure-As-A-Service. https://www.redhat.com/en/topics/cloud-computing/iaas-vs-paa...

The people who don't see the value in generating language that has a purpose outside of narrow niche of communicating facts will be let down for some time. This feels very Wittgenstein's Tractatus. There are so many other ways that we use language.

for sure!

My grinder is a Weber EG-1. It has really large (80mm) flat burrs. Very forgiving; like I don't have to be dialed in perfectly for the espresso to be very good. It's essentially a cafe quality grinder, but optimized for a single-dose flow.

My machine is an Olympia Cremina, which is a manual lever machine. I've modified it with a pressure profiler to control and track the pressure of the shot I'm pulling. All that manual control is really nice for the kind of coffees I like which are usually naturally-processed light roasts. However, manual levers are not for everyone and have a definite learning curve.

If you're curious about either of those, James Hoffmann has good reviews of them both on YT.

I really enjoy the nerdiness and tactile feedback of the manual lever, but if I didn't want go manual, I'd look at a Decent DE-1.

If I had to shave a bit off that budget, I think I'd go with a Rancilio Silvia Pro X, which is half the cost.

If you do get upgraditis at any point, I'd definitely consider swapping that priority. I think, I'd rather use high-end grinder with a mid-tier machine than the other way around.

You can spend more but the coffee doesn’t get better

I went down the espresso rabbit hole during covid, as did a good buddy of mine. We both upgraded from setups in the ~$1k range and are experiencing much better results in the cup. Obviously you'll see diminishing returns at some price point, but most of my friends/family agree that a latte from my new setup is noticeably better than one from my old setup.

I strongly disagree. And not just because copyright in perpetuity is unconstitutional. The value in a rich public domain is vastly under-appreciated. The default position is that IP is not protectable by law, because the free exchange of ideas is extremely important to modern society.

Certainly we have carved out exceptions to that default position, but only for very clear and distinct policy reasons. 1. consumer protection (trademark) and 2. incentivizing innovation and expression (patent and copyright).

The idea that my great-great-grandchildren might want to benefit from my having written a book really does not factor into whether I might write a book. If I'm not incentivized by life of the author + 70 years, I would probably not otherwise be incentivized.

Also, corporations don't pay to extend their copyrights. Other than the money Disney pays their lobbyists.

I know essentially zero about this space, but I would almost argue categorically that 20% is not "incremental".

Media is a big issue. MMS is potato quality compared to what you can send over RCS.

I like high grain images, but I also wonder if it's kind of a fad?

The ability of modern digital cameras to shoot essentially unlimited photos might not be without its drawbacks, but being able to shoot low grain images practically in the dark is pretty cool.

I spend a fair amount of money on coffee, but I know what styles and roast levels I like, and I know roasters that do those things well. If you like weird varietals processed with some experimental anaerobic processing method and a light roast, Onyx and Black & White, for example, both do really well with those styles. And they aren't cheap, but I enjoy them. If you favor medium to dark roasts, then Dunkin Donuts brand coffee in the grocery is pretty darn good.

Point 1 is very similar in high end coffee. Sometimes very expensive coffee in the US is substantially better than what you get in the grocery store, but often it's just much different. It might taste like blueberries or strawberry candy or wine. Not necessary better, just weird, hopefully in a whimsical way.