You need to use the features that Claude Code gives you in order to be successful with it. Your build and tests should be in a Stop hook that prevent Claude from stopping if the build or tests fail. Combining this with a Stop hook that bails out if the first hook failed n times already prevents infinite loops.
With anything above a toy project, you need to be really good with context window management. Usually this means using subagents and scoping prompts correctly by placing the CLAUDE.md files next to the relevant code. Your main conversation's context window usage should pretty much never be above 50%. Use the /clear command between unrelated tasks. Consider if recurring sequences of tool calls could be unified into a single skill.
Instead of sending instructions to the agent straight away, try planning with it and prompting it to ask your questions about your plan. The planning phase is a good place to give Claude more space to think with "think > think hard > ultrathink". If you are still struggling with the agent not complying, try adding emplasis with "YOU MUST" or "IMPORTANT".
I'm not an astronomer, but my intuition is this:
When the source of the light is moving towards the observer, each successive photon emission happens from a position closer to the observer than the previous photon. Hence, from the observer's perspective, the time between photons is reduced, meaning more photons are observed in a given time, and the brightness is increased.
When we observe a galaxy that is rotating opposite to us, not only is the source of the light moving closer to us, but we are also moving closer to it.
From the perspective of us looking directly down on a parallel “plate” of a galaxy (with the other galaxy viewing us in the same way), relative differences in speed for the situation of same direction of rotation will be much smaller than for opposite directions.
But between any point in one galaxy to another, just as much matter will be moving closer as moving away. Regardless of same or opposite rotations.
But perhaps greater red shift and greater blue shift (as apposed to lesser of both) as a practical matter of telescopes vs. their cross spectrum sensitivities, means more light detected.
When it gets cold, the homeless congregate in the warm interiors of malls. The guards on duty won’t let them sleep there, but they prefer it over being out in the cold.
Yeah, it's definitely not a perfect solution. Portland, OR has tried a housing first policy. However, there weren't enough available houses, so that requires to build them. That's lead to years of people still being homeless and underserved while resources went into building homes.
I think it can help a lot under the right circumstances, but if the system is already overwhelmed then funding less permanent solutions for a while may actually be more effective and kind
Building homes does not take very long if the regulations allow it. Most of the US needs to take a hard look at regulations. Often the problem is regulations won't allow a small house and so the homeless are forced on the street because they can't afford anything.
Onerous regulations and process are a huge part of the problem. I looked into building a small apartment complex with a local developer in the city. It was zoned well, but the entire surrounding neighborhood had dirt roads. After talking with the city planner we found out that:
1. We'd need to pay 5k get a price estimate for all of the permits we would need, and
2. That we would need to either pave a road along the property (which would immediately fall apart from erosion since it would be the only paved road for multiple blocks), or pay $400k for the right to not pave the road
At that point, we stopped even trying. There's no way to build affordable housing in a timeline manner in that city.
Often the regulations are the start of the problem.
I remember reading an article in Portland that the city was expecting N new people to move there in the next 10 years. So they approved construction of housing for N-200k people.
They intentionally set things up to push the poorest and most vulnerable 200k people into homelessness.
How come a fully typed ORM is the devil, if we agree we want a typesafe codebase for our mixed experience dev team?
I have had positive experiences with Prisma. It just works.
I personally draw a distinction between micro-ORMs and ORMs. A micro-orm will simply take a strongly typed flat struct and map it into the set of parameters for a query, and likewise map a single row out to another flat struct (or enumerate/iterate while doing so). They may even include insert, update, and delete helpers that deal with a single table. This is what 90% of Prisma does and is the Good Parts. Migration generation is also good (but can be dangerous, e.g. deleting columns).
I'll call out query generation separately, as it is a lesser evil (in my opinion). This falls under a larger peeve of mine, which is using data (JSON, YAML, or this[1]) as programming languages. Using data as a programming language sucks. Tooling (compilers, LSP, etc.) are typically absent, meaning that mistakes very much become a runtime issue (not really a problem for Prisma). The deeper problem is that you'll run into limitations and will: either have to drop down to SQL (tricky given that you've rarely used it thanks to using ORMs to generate all your queries), or kludge/hack it up in order to remain in the ORM land. There's also lines of code and readability to contend with, the second Prisma example (the first is on my shit list for reasons further down):
SELECT u.email FROM users u
WHERE (u.email LIKE '%prisma.io' OR u.email LIKE '%gmail.com')
AND u.email NOT LIKE '%hotmail.com'
Just write the fucking SQL.
The objectively bad parts are one or more of the following:
* Change tracking: magically being able to update a value returned by the ORM and calling `save` to write it back to the DB.
* Object graphs: magically accessing related objects in memory, e.g. `order.orderlines` or `order.address.city.state`.
Relational databases (SQL) are not graph databases (in-memory object hierarchies). They are orthogonal concepts. Just throwing whatever you have as your classes into your database is neglecting to think about how your data is stored. This will come back to haunt you. You might claim that you can do the careful design using classes, and I would believe you, however, those juniors you mentioned aren't going to have the knowhow (because they've been shielded from SQL by ORMs their entire career and so don't understand how to make good databases).
Case in-point, back to that Prisma example shown earlier. Any ideas what's actually wrong with that query, besides the pointless hotmail check? Both the original and my conversion have the same serious issue that stems from not designing the database.
The fashionable thing for a while now has been to write your react webapp as a single app.
Here is one example of a "fullstack" react template: https://create.t3.gg/
Interesting setup. Looks like getting to a zen state like this would have required them to migrate their backend from Rails to ts at some point. I can understand the engineering decision not to do that.
Worse than not trying is trying and experiencing burnout and/or destitution.
When you fail, it can be due to many things. Not everything in the world is controllable. This is one of the reasons why expecting zero ”What ifs” at the end of your post-mortem is unreasonable.
I don't think there's any absolute rule here. I'm pretty convinced that a startup would be healthier for me rather than my safe day job. The amount of churn and nascent burn out due to chasing tech debt and tickets surrounded by demotivated people is really bad.
I guess being aware of what you need and can afford is key.
It's fine in my experience. I use Synergy 1 and I turn off smooth scrolling on my laptops and desktops. I find it more distracting that some apps support smooth scrolling and some don't, so I just turn it off altogether.
Together with AutoHotkey on Windows and (Built-in) Applescript and Rectangle on MacOS, one is able to have similar keybinds on both operating systems.
Every app on macOS supports smooth scrolling. That’s why people use it. You’re using the lowest common denominator and going “wow everything is the same I can’t tell the difference”.
Only with hardware acceleration turned off. Which in my experience causes a stuttery and laggy experience which can hardly be described as "smooth".
If you really want to use MacOS's smooth scrolling, use MOS (https://mos.caldis.me) and get a beefy Mac because the base models will stutter like I mentioned.
I am very confused what you are talking about. You definitely do not need hardware acceleration to be disabled to get smooth scrolling. Something seems very broken with your setup?
4 years ago I was 100kg @ 185cm. Never touched PEDs. All it took was being in the gym going all out for around 2 hours, 6 days a week. My body fat was 13% consistently. I tried going below 13% until it started to affect my strength.
The rest of my life also revolved around recovery and nutrition because very quickly I realized that I needed to make extraordinary efforts to allow my body to recover from the regimen. I slept 10+ hours and ate around 3300 cals daily (TDEE was massive).
Being on antidepressants at the time made eating these amounts much easier, as it seemed to turn off something in my brain responsible for me feeling satiated.
I started gym after being diagnosed with the psych issues. I really put 110% effort into it and got… big stretch marks because my muscles grew so fast. I don’t think what I was doing was ”healthy” as much as it was ”I am super depressed and heavy barbell squats is my coping mechanism”.
I just wanted to say, you don’t need PEDs to get ”scary” jacked. But you probably need to be crazy like me.
If (and that is a big if. Most people severely underestimate their body fat levels. And with some touting 3% dexa scan results, even such measurements are suspect) you really were at a bmi of 29 (which, BTW, doesn't count as 30+ from my original statement ), then congrats on your amazing genetics and work ethic.
Still, to break that bmi 30 barrier significantly, while maintaining the same level of leanness, you'd likely require some additional 'help'.
And again - it is possible for a number of gifted people, just highly unlikely. Most males can't get that jacked no matter how much and how hard they train. I'd say that even with PEDs most wouldn't break that barrier while staying lean.
My personal best was a 28 bmi with what looked like a 9% body fat, veins fully visible at the lower abs and the hips etc, and it was absolutely unsustainable for more than a year, for a plethora of reasons. Can't even imagine gaining 3 more points.
To reiterate: I'd still focus on a BMI to gauge the public health - the genetic outliers are rare
Everyone hates hearing this one:
Documentation, documentation, documentation.
Programming is a social task. Therefore, everything else related to software development best practices branches off from that.
With anything above a toy project, you need to be really good with context window management. Usually this means using subagents and scoping prompts correctly by placing the CLAUDE.md files next to the relevant code. Your main conversation's context window usage should pretty much never be above 50%. Use the /clear command between unrelated tasks. Consider if recurring sequences of tool calls could be unified into a single skill.
Instead of sending instructions to the agent straight away, try planning with it and prompting it to ask your questions about your plan. The planning phase is a good place to give Claude more space to think with "think > think hard > ultrathink". If you are still struggling with the agent not complying, try adding emplasis with "YOU MUST" or "IMPORTANT".