Nathan Van Gheem

2.7K posts

Nathan Van Gheem banner
Nathan Van Gheem

Nathan Van Gheem

@vangheezy

The curious task of "Engineering" is to demonstrate to men how little they really know about what they imagine they can design.

Raleigh, NC Katılım Temmuz 2009
306 Takip Edilen570 Takipçiler
Nathan Van Gheem retweetledi
David Cramer
David Cramer@zeeg·
people like to think this (especially indies), but the reality is saas is less about the tech, and more about liability reduction folks have been able to self-host sentry for all of time, why don't they? nothing changes- bad businesses continue to be bad businesses
Cody Schneider@codyschneider

had dinner with engineer lead at a startup last night told us their saas vendor tried to double the price from $70k a year to $170k a year CTO was like in slack on the call "i think we can have claude write this" 3 week sprint cloned and replaced vendor told sales org trying to gouge them they done sales freaks out "how do we solve this" this is going to happen so much in 2026

English
41
21
506
72.1K
ThePrimeagen
ThePrimeagen@ThePrimeagen·
after watching my son attempt to vibe code for ~15 hours, I am more convinced than ever hard skills are going to be needed
English
188
109
4.7K
242.1K
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@zeeg is sentry going to eventually integrate with infrastructure/system metrics? The product is so good with errors and traces right now. If only we could get system metrics integrated, I wouldn't need other solutions.
English
1
0
0
15
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@moo9000 I imagine they had to pay it, they just didn't collect it. There's no way they didn't pay.
English
1
0
1
40
Mikko Ohtamaa
Mikko Ohtamaa@moo9000·
On "European mind cannot comprehend" bit. What we do not comprehend - it's bad that companies violate rights and laws, and the companies should be punished? Vercel is net positive for the US economy. There might have been lax sales tax collection at the start, which themselves are not very significant, if I understand correctly. This has now been paid back multiple times, in orders of magnitude. By having sustainable growth and not focusing on fighting with the government has helped Vercel to create many more job opportunities and thus employ more people. Such employment creation opportunities are harder to come by in the EU, because businesses need to focus more on fighting the bureaucracy, much of this being non-productive work bringing very little value to society, and this is one of the reasons the EU has stagnated for the last twenty years. Of course, any violations must be fixed long-term, but focusing on violation preventive regulation at the cost of productivity in the short term is away from growth, and if you do not have growth, then you do not usually have long-term in the first place.
Adam Wathan@adamwathan

Founded 10 years ago, $9B valuation, will only start worrying about sales tax next month. The European mind cannot comprehend.

English
9
1
20
3K
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@mitsuhiko I agree with this so much. Even apart from dep maintenance hell, too many engineers think deps are "free" from risk. You introduce that dep, now you have to support it in prod. If something goes wrong, now you're debugging more layers that you could have written yourself.
English
0
0
2
319
Nathan Van Gheem retweetledi
Armin Ronacher ⇌
Armin Ronacher ⇌@mitsuhiko·
“Celebrated are the minimal dependencies, the humble function that just quietly does the job, the code that doesn't need to be touched for years because it was done right once.” lucumr.pocoo.org/2025/1/24/buil…
English
16
47
297
27.9K
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@domenkozar @dnouri I don't think tokio calls them that. Async is single threaded and better to not use the word "thread" to describe the execution model probably?
English
2
0
0
30
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
it's 2024 and argocd still shows "Failed to load data" while loading the data and eventually showing the data
Nathan Van Gheem tweet media
English
0
0
0
112
pfreixes
pfreixes@pfreixes·
@DHatanian Which ones? raw material? or also others that are intrinsic of NVIDIA? but wouldnt be still a good call?
English
1
0
0
74
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@pfreixes Right, maybe not so different from a monorepo in some cases. In the end, this is software development and it's all about physical/logical boundaries. At least with "modular monolith", maybe it will prevent people thinking that every logical boundary needed to also be physical.
English
0
0
0
25
pfreixes
pfreixes@pfreixes·
Im gonna need to read it, but modular monolithics smells a lot to microservices but with the win on having it in the same code base
Dr Milan Milanović@milan_milanovic

𝗪𝗵𝘆 𝗱𝗼𝗲𝘀 𝗚𝗼𝗼𝗴𝗹𝗲 𝗿𝗲𝗰𝗼𝗺𝗺𝗲𝗻𝗱 𝗠𝗼𝗱𝘂𝗹𝗮𝗿 𝗠𝗼𝗻𝗼𝗹𝗶𝘁𝗵𝘀 𝗶𝗻𝘀𝘁𝗲𝗮𝗱 𝗼𝗳 𝗠𝗶𝗰𝗿𝗼𝘀𝗲𝗿𝘃𝗶𝗰𝗲𝘀? In the last decade, we have seen a massive trend of using microservices everywhere. We were building systems for a few hundred or thousand users and wanted to know how to make a system for millions of users. This was over-engineering and needed to be corrected. Why it was wrong? Because the development lasted long and we created incredibly complex systems, hard to maintain. This is especially true for startups that must go fast and stay simple. A recent paper by authors from Google found that most of their developers split binaries for one of the following reasons: it improves performance, fault tolerance, and abstraction boundaries and allows for flexible rollouts. Yet, splitting applications into microservices has its challenges: 🔸 𝗜𝘁 𝗵𝘂𝗿𝘁𝘀 𝗽𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲. The overhead of serializing data and sending it across the network is increasingly becoming a bottleneck 🔸 𝗜𝘁 𝗵𝘂𝗿𝘁𝘀 𝗰𝗼𝗿𝗿𝗲𝗰𝘁𝗻𝗲𝘀𝘀. It is incredibly challenging to reason about the interactions between every deployed version of every microservice. 🔸 𝗜𝘁 𝘁𝗮𝗸𝗲𝘀 𝘄𝗼𝗿𝗸 𝘁𝗼 𝗺𝗮𝗻𝗮𝗴𝗲. Rather than having a single bi-nary to build, test, and deploy, developers must manage 𝑛 different binaries, each on their release schedule. 🔸 𝗜𝘁 𝗳𝗿𝗲𝗲𝘇𝗲𝘀 𝗔𝗣𝗜𝘀. Once a microservice establishes an API, it becomes easier to change by breaking the other services that consume the API. So, they proposed the following approach: 𝟭. 𝗪𝗿𝗶𝘁𝗲 𝗺𝗼𝗻𝗼𝗹𝗶𝘁𝗵𝗶𝗰 𝗮𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 that are modularized into logically distinct components. A component is a long-lived agent, similar to an actor. 𝟮. 𝗟𝗲𝘃𝗲𝗿𝗮𝗴𝗲 𝗮 𝗿𝘂𝗻𝘁𝗶𝗺𝗲 𝘁𝗼 𝗱𝘆𝗻𝗮𝗺𝗶𝗰𝗮𝗹𝗹𝘆 and automatically assign logistical components to physical processes based on execution characteristics. So, if both components are in the same OS process, they are called regular method calls, but if they are co-located, calls are executed as RPCs over the network. Runtime decides whether these modules should be collocated or moved to different machines (and scaled, etc.). 𝟯. 𝗗𝗲𝗽𝗹𝗼𝘆 𝗮𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗮𝘁𝗼𝗺𝗶𝗰𝗮𝗹𝗹𝘆, preventing different versions of an application from interacting. This approach consists of two main parts: a programming model with abstraction that allows developers to write modularized applications and a runtime for building, deploying, and optimizing these applications. They claim that it reduces application latency by up to 15x and costs by up to 9x by simplifying application management and deployment. If you want to check the framework implementing the approach from the paper, check https:// serviceweaver. dev/. What do you think about this approach? Does it look like EJBs or CORBA? #microservices

English
2
0
1
105
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@mitsuhiko @dhh it's good to have someone test against orthodoxy at the very least -- show some alternative paths The amount of hate he gets for benign opinions is wild though
English
0
0
1
1.1K
Armin Ronacher ⇌
Armin Ronacher ⇌@mitsuhiko·
Y’all might think @dhh knows nothing and is needlessly contrarian but I think there is something behind both cloud avoidance and single tenancy for databases. The time feels right.
English
22
11
284
78.1K
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@fulmicoton @antirez It's unfair to pick and choose the cases one language behaves better then another I do think Ruby has some nice designs but Ruby vs Python is boring
English
1
0
0
174
Paul Masurel 🦀
Paul Masurel 🦀@fulmicoton·
(following the @antirez rant) All programming language come with footguns. What's your favorite one? Here I go:
English
4
0
4
2K
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
"let me chatgpt that for you" doesn't roll off the tongue like "let me google that for you"
English
0
0
0
115
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@mauricioaniche I think reality is closer to the fact that people are too lazy to learn the real db layer, write sql and data mappers and not that they are too lazy to understand an ORM
English
0
0
0
90
Maurício Aniche
Maurício Aniche@mauricioaniche·
People say “our queries are too complex for a ORM”, or “ORMs are too magical”. What I hear is “I don’t really know how to use one and I’m too lazy to learn” For most apps out there, you need a reason NOT to use an ORM, and not the other way around.
Gavin King ⍼⍼⍼@1ovthafew

If anyone is truly still confused about how to to relational data access using ORM in the year of our lord 2023, I'm begging you to click the following link and start reading from there: #association-fetching" target="_blank" rel="nofollow noopener">docs.jboss.org/hibernate/orm/…

English
16
6
85
23K
Nathan Van Gheem
Nathan Van Gheem@vangheezy·
@mauricioaniche I've known many ORMs well, written my own sort of one but now I recommend against them because they usually aren't necessary and just add an unnecessary layer. A good middle ground is something like python sqlalchemy's table/mapping layer BUT NOT the ORM
English
1
0
4
279