Unit tests are just as important as integration tests as long as they're tightly scoped to business logic and aren't written just to improve coverage. Anything can be done badly, especially if it is quantified and used as a metric of success (Goodhart's law applies).
Integration tests can be just as bad in this regard. They can be flakey and take hours, give you a false sense of security and not even address the complexity of the business domain.
I've seen people argue against unit tests because they force you to decompose your system into discrete pieces. I hope that's not the core concern here becuase a well decomposed system is easier to maintain and extend as well as write unit tests for.
The problem with unit tests these days is that AI writes them entirely and does a great job at it. That defeats the purpose of unit tests in the first place since the human doesnt have the patience to review the reams of over-mocked test-code produced by AI.
The end-result of this are things like the code leak of claude code presumably caused by ai generated ci/cd packaging code nobody bothered to review since the attitude is: who reviews test or ci/cd code ? If they break big-deal, ai will fix it.
the problem is people make units too small. A unit is not an isolated class or function. (It can be but usually isn't) a unit is one of those boxes you see on those architecture diagrams.
Inability to unit test is usually either a symptom of poor system structure (e.g. components are inappropriately coupled) or an attempt to shoehorn testing into the wrong spot.
If you find yourself trying to test a piece of code and it’s an unreasonable effort, try moving up a level. The “unit” you’re testing might be the wrong granularity. If you can’t test a level up, then it’s probably that your code is bad and you don’t have units. You have a blob.
> mozilla is a non-profit so doesn't need to respond to market signals.
No? They, just like every other entity, will have to achieve it's investors goals using the least amount of costs possible.
It just happens to be the case that Mozilla's investor's goals aren't more money.
For example, if a non-profit wants to build a bridge to a small island to provide the children of that island access to the mainland's schools; that non-profit would still be very susceptible to market signals regarding the most cost-efficient materials for bridge-building.
It is precisely because of current market signals that this move is a bad move for Mozilla. If GPUs were a dime a dozen and an AI engineer cost a thousand times less to employ than a browser engineer, Mozilla offering an LLM service would be a lot less objectionable than in the current economy.
> No? They, just like every other entity, will have to achieve it's investors goals using the least amount of costs possible.
The Mozilla Foundation is funded by donors not investors. The fiduciary duties of non-profit directors do not have to include using the least amount of costs possible.
The Mozilla Foundation's stated goals are "to advance the vision of the future of the internet and technology".
You might reasonably argue that this means copy catting every other product by slapping AI on itself, but I would counter that this actually demonstrates a lack of vision.
FF is still a solid choice, privacy-wise (with some manual tweaking), but just in the past few years 80-90% of their revenue came from adtech partnerships, so expect a series of rug pulls like the recent ones.
Being a non-profit that does not have a share-price, I'm curious if you have evidence for validity of your assessment of "big swings resulting in compensation packages".
I can offer a rationale that I hope that will be interesting. Evidence is more effort than I'm willing to invest in a random internet conversation. I hope that's okay. I don't mean to cause offense, but neither do I like having an interesting discussion shutdown.
Executive leadership compensation tends to be based on,
- the size/prospects/complexity of the company
- the compensation received by executives in similar roles at other companies
- the amount/type of oversight by the board
This incentivizes executives to increase the complexity of their role in order to justify greater remuneration. The classic example is turning a widget factory into a widget financial services provider. In this case, by behaving like Silicon Valley companies chasing the latest fad the executive leadership of Mozilla get to demand the same remuneration packages.
Those are false equivalencies. The EU bureaucracy is a mess of competing factions. It isn't a corporation or a political party. A better analogy would be a US Government department breaking a law.
Ideally there are are legal avenues for redress in your situation. If there are not, that is more likely to be the problem than a conspiracy of propaganda.
I agree with you, but this is what happens in reality. Laws and regulations get abused more than they get used, unless they are very well written. The minimum would be to require the EU itself to not use Microsoft 365 and other services that they have explicitly tried to counteract with their laws.
Change the word "propaganda" to "marketing", if it is offensive to you.
You get ads outside the UK, which is always weird for me when i get back from holiday and have a bunch of shows i downloaded while I was in another country and haven't listened to yet. Actually, I say bunch - but I've deleted everything except In Our Time. Australian Broadcast Company does more better podcasts than BBC these days.
The BBC is another bad actor in this space (although increasingly irrelevant). Their decision to "hide" most of their RSS feeds whilst still labeling their proprietary app subscribe links podcasts was unethical.
I see hope in the Patreon model. I don't mind paying a monthly subscription for a specific show if I get a locked down RSS link.
Though, podcasters on Patreon are getting high on their own supply, asking for way, waaaay too much. "Just" 5 bucks a month? For one show??? Completely out of touch. Come back when it's ad-free @ 50c, or less.
I dropped the two BBC podcasts i'd been interested in (News Cast and In Our Time) for this reason, it's such a weird tack to take in order to chase views.
I support a few podcasts on patreon. They have regular open RSS feeds and then a locked down patreon feed, containing special episodes, extra material etc. One podcast only charges for each locked episode, so if they don't make any for a while I'm not charged. They are all pretty niche and I don't know how much they earn from it, but from a user perspective it's working fine.
I'm trying to understand what you mean by this. Their podcasts still show up in e.g. Apple Podcasts and Pocketcasts. If you look at the 'podcasts' page for e.g. '13 Minutes to the Moon'[1] (which I hadn't realized there was a third season of), there is a (small) link to the rss[2]
You think even ten years from now, much less 1,000 years from now, whatever LLMs turn into won’t be at least as capable as the best human of following code paths?
We can spin up a million of them and run them at 10,000x speed.
Integration tests can be just as bad in this regard. They can be flakey and take hours, give you a false sense of security and not even address the complexity of the business domain.
I've seen people argue against unit tests because they force you to decompose your system into discrete pieces. I hope that's not the core concern here becuase a well decomposed system is easier to maintain and extend as well as write unit tests for.