mattbrewsbytes

There's no way getting around needing to install the patched versions of the compiler or runtime. Someone has to do it and in my opinion it should be done internally so its controlled and has verified/known sources.

I think option/bullet 2 keeps things as streamlined as possible - base standard image for java apps that apps then use as their base when they build their container images to run in k8s containers. Maybe there's a way to tag base images with major.minor so that apps just pull that image and if its cached its faster and if there was a major.minor.patch just applied and you're the first app to deploy then that app pays the penalty of waiting for the underlying base image to be rebuilt.

A volume is an interesting idea but are there possible issues for OS library dependencies like openssl, database drivers, etc. that might not be the right version for a JDK mounted on a shared volume?

It doesn't matter what knowledge based industry it is either, its always people, communication and decision making (or lack of) that makes everything take longer.

Get a committee together to decide multiple products priorities, features, designs and you could be months away from having anything defined enough to code.

What gave you the impression during what I would assume is a 1-2 hour interview that the person applying for a job is foreign agent?

Do the behaviors exhibited also align with other somewhat common candidate behavior like not being competent in what they say, being nervous, etc.?

Meta can't scan my phone if I don't install Meta's apps on my phone.

A web browser on the phone removes the need for a lot of "apps".

Its fun watching things swing back and forth over time. I remember having those Sun mini-fridge size servers, all running RISC sparc based CPU's if I remember correctly. I wonder if there would be some merit in RISC based linux servers, like maybe the power consumption is lower? I forget the pros/cons of RISC vs CISC CPUs.

Syntax of things I don't remember or care about, like regexes in language X or Y. Syntax of things I used to know but I'm not up to speed on the latest versions. For example, using tailwind, I can just describe in english what I want to happen with page elements and for the most part AI gets the code right, sometimes its wrong and I need to debug.

A common argument I hear about AI is "I could just write it faster myself", well I know CompSci and general info about a lot of software things but it would take hours of getting up to speed on areas I'm not an expert in to be productive. I can just delegate that to AI and get mostly correct outputs, this is okay with me and faster than what I could do.

I think the cost is going to catch up with the AI companies running the models (not the companies building products that call AI APIs) and that is when the bubble will burst. They will need to keep increasing costs and at some threshold, fewer and fewer developers in an organization will have licenses because it may become unaffordable.

I think it comes down to quality of materials and manufacturing. It would be interesting if someone knew of failure rate data of laptop manufacturers. I bet Apple's is far lower than others, maybe this is because they control the OS and firmware, etc. I don't care enough to go research but an HN thread isn't likely representative of "normal" laptop users.

I'm not an Apple fanboi, I have a lot of linux experience, back in the day built a Tivo like linux PC with TV capture card as a DVR and had to mess with all sorts of X11 settings, etc. Used to build PCs for gaming and mess with settings. The whole It Just Works is true with Apple, everything hardware wise is smooth. The annoyances of Apple software to me don't bubble up to the level of wanting to switch, but articles like above make me think about getting a cheap Mini PC to play around and see if things like Omarchy make the level of messing with settings much lower than it used to be.

My rationale given on another comment is when writing web apps, everything comes in as text so you have to parse/convert and validate parameters anyhow.

Dynamic makes a lot of sense for web programming. Static should definitely be used in high risk software where its life and death, medical devices, rockets, nuclear plants, etc.

If you’re checking and validating inputs into a method and you’re writing web applications where everything is text over HTTP, having type checking notations, etc. are a bit overkill.

Cisco still exists and was a winner for a long time. The company that emerged hugely successful out of the dot com era was Amazon, they dominate ecommerce today.

Take a look further back to the Personal Computer era in the 80's into early 90's. There were tons of PC makers out there, IBM became dominant but they haven't been in that business for a long time now. Microsoft was the enormous winner coming out of that.

When grandma starts talking about GPU's and NVIDIA stock, this AI bubble might be at its peak. I think if history rhymes with past lyrics (that whole history doesn't repeat but rhymes thing), its more likely that a software company (OpenAI, Google, Microsoft) is going to be the dominant player emerging from a potential bubble burst.

A potential pin-prick to the bubble (looking into my crystal ball): pricing for AI - once companies realize their runways are running out and they want to turn a profit of some kind to stakeholders they will need to raise prices to offset the energy and hardware build outs. Customers may realize they aren't getting the productivity to offset those increased costs and hold back budgets on AI spend. Or chip tariffs depending on TACO volatility.

The key thing no LLM or AI is capable of is innovation. The basis of all their inputs came from humans. If a company needs to pivot, good leader(s) can execute that and have tremendous impact, I wouldn't trust AI to come up with that strategy.

The reason AI/LLMs are talked about augmenting/replacing people who code is because LLMs are really good at well defined, structured text output and there is a large amount of training already done. Let the LLM/AI do the tedious, repetitive parts of the work and let the engineer spend their energy on harder problems.

Pattern recognition. I don't mean the ability to do regular expressions. I mean the ability to recognize patterns in software development, organizations, and trends in the industry. It helps recognize when you've seen something similar before and know a good approach to dealing with it. This would include design patterns (like Gang of Four), integration patterns, architecture patterns, etc. There are a lot of problems out there that have stable, known solutions to them.

One part of maintainability is cost. And a simpler mechanical engine without proprietary ECUs is going to be cheaper to maintain, provided parts are available.

If someone encounters issues with an ECU and it needs replacement at $1k-2k they might just consider the costs and that being a down payment on a new vehicle vs. repairing. Labor costs more than parts for complicated electrical/computer/engine problems. Electrical issues in modern vehicles don't appear to be easy to troubleshoot, sometimes require proprietary tools. A simpler mechanical engine could be DIY repaired and running, check out the "low-buck garage" youtube channel and the $2 Jeep series as an example.

I'm not advocating something like going back to computer-less, inefficient, stinky, loud cars, just pointing out that when we add computers to things, it makes them less maintainable to the average person.

Learning n8n a little more for some automated tasks on a hobby project. Maybe try re-learning sketchup to diagram some home improvement/woodworking projects.

Probably overthink buying a mini PC to run Ubuntu on to do homelab like things, like running n8n or other docker images on like plex. Then probably abandon the idea and watch something off a streaming service, lol.

I don't see a question in there, this feels more like a blog post and not AskHN.

I don't think anything with mechanical moving parts is going to last that long, with regular usage, and have original parts.

The fact that the owner can keep it going is a testament to the maintainability of combustion engines that don't have high tech computers in them.

When do specific actions taken by the current US administration cross the legal line into treason or other legal lines where defense against domestic enemies is warranted?

I’ve think the “call” part is a bit from the past, should be a setup account online with Vanguard, put it in VOO or VTSAX or the equivalent low fee market index fund and leave it alone.

Replace Vanguard with any other firm but the key is picking low fee ETFs and leaving them alone. Vanguard tends to have a reputation for the lowest fees.

What you have stated is almost the same as call Vanguard, all those options are the same in the sense that they all involve investing, leaving it alone for a long time. Its just the vehicle thats slightly different and tax advantages.

I wouldn’t consider those options needing much motivation or research. The key with all of them is investing early and leaving it alone.

A slightly different viewpoint is that sharing in public or larger private channels allows for knowledge sharing and collaboration. Sometimes the key person is wrong because they aren't the only one working on something. I know that ego might get in people's way sometimes but other people in the team and in the organization also have valid perspectives. As a manager, its important to try and get to a best solution and that means collaboration, not a specific person's approach all the time.

The redundancy also helps the key person be able to disconnect when on vacation. If you are the sole knowledge base for some critical part of the company, might as well drag the work laptop with you every where you go.

You didn’t explain why you wanted to use debian, you don’t have to, but there are a lot of x86 based laptops out there that will run various linux distros so I wonder why you’re on a Mac to start with? Employer choice and not yours?

You mention Typescript which I’ll assume means some sort of web development, one thing to maybe consider is looking at https://omakub.org/ for inspiration on hardware choices and a setup. I don’t use it but was looking at it recently out of curiousity (I do Rails so thats how I’m aware of DHH and him leaving Apple HW).

If you work at a place that has a modern CI/CD pipeline then your multiple deployments per day are likely rebuilding that large bundle of JS on deploy and invalidating any cache.

HTTP 2 has been adopted by browsers for like 10 years now and its multiplexing makes packaging large single bundles of JS irrelevant. SPA’s that use packaging of large bundles doesn’t leverage modern browser and server capabilities.

General investment advice is to put like 90% (or more) into ETFs, index funds or low-fee mutual funds. That last 5-10% could be individual stocks if you want to play around and do research, etc. If someone went all-in on AI companies and a stock market bubble bursts, you're in the "find out" phase.

Regarding bubble bursting, there have been market protections put in after 2008/09 so I don't know that we'll see major stock crashes. Its more likely that companies may miss targets, products start plateauing on features, etc.

If a crash were to happen, I read in a book (Intelligent Investor maybe?) there are usually macro indicators of patterns to look for, I forget what they are but its things like GDP, jobs, CPI, major indices falling X out of Y weeks in a row and other things like that. But that was all from a book looking in hindsight at prior crashes, prior to 2008 and whatever changes they've made since then.

Your own definition seems to contradict itself. Slack is horizontal, like Teams, send files, real time chat, video huddles, channels based on access, integral with other tools via plugins.

Not every tool is going to be HIPAA compliant, or the equivalent in other countries.

Can someone explain horizontal vs. vertical SaaS? This is new to me and I can't figure out what this means... its all SaaS so why differentiate? Seems like just marketing buzzwords.

> automating software development for ad-tech companies

A lot of technology advancements are automation of human tasks. Its been going on for decades and does eliminate a lot of jobs (or move them to a different continent). There aren’t telephone switch operators anymore, or cashiers adding up items manually in a store checkout or any other countless jobs that are now done with software.

One way to look at it is if you’re an employee at a company you don’t really have any say in what projects/products you work on. If those projects are eliminating or creating jobs or saving lives or whatever, your only choice is really about working for that employer.

Workers never benefit from productivity gains, doesn’t matter what is fueling those gains, AI, factory technology, automobiles instead of horses, etc. Business owners and executives benefit because they are at the top of the hierarchy where the profits are scooped.

Sure there are workers that get bonuses, but your bonus for a worker or lower level manager typically is pre-determined so if you overwork and contribute 50% more productivity enhancing output, you don’t get 50% more in your bonus.

One path could be to build your own thing and become a business owner but that is much more work than just building software.

Do companies like Oracle and SAP run their own business on their ERP platforms? That might be the biggest one.

The fact that this was discovered by a novel satellite processor makes an uneducated person like me on this topic wonder if this has happened before and us humans just didn't have any way of observing it. We've only had satellites for a few decades and earth and its oceans have been here a while longer than that.

I've been using LoseIt and I've had a successful six months this year and realized a key thing: linking to Apple Fitness was slowing my progress. Here's why - the app tracks "workout calories" at a very high rate and it was essentially calculating my daily calorie budget back to maintenance calories, i.e. not losing and not gaining. I unlinked Apple Fitness from the app. Check the math from your app, use a TDEE calculator as a reference, a lot of apps and fitness devices over estimate calories burned.

I just realized I follow a single responsibility principle for apps: LoseIt for calorie tracking, Apple Fitness for tracking exercise and Streaks for tracking habits. They're all a tap or two away, no need to combine them for me.

Regarding custom recipes - I've found the search results in LoseIt have enough results that I can judge what I ate and pick something pretty close. Calories on packages of food or published by restaurants are working with estimated averages anyways so my opinion is being exact isn't going to result in more weight loss, for me it will result in frustration and likely I stop tracking. If I'm +/- 100-200 calories for the day, I'm still within the range of losing and this fits more with making changes I can stick with. Not saying others should adopt this approach, just sharing my thinking on it, you do you.