> Did we learn nothing from the TDD nonsense in the 00s?
I was never a fan of TDD, but ironically I think it may finally be a good development strategy now with agentic/LLM coding since LLMs do so much better they they have a verifiable goal like passing tests.
How does that work? Does one agent write the tests from spec, and then another agent is told "make tests pass", or do you write the tests manually?
It does sound like a possible way to parallelize agentic work, but I'd be surprised if the results are better than a non-TDD approach where the agent would write both tests and code from spec.
I would expect the agent to suffer the same problems humans do with TDD, namely that it is hard or impossible to properly model data ("understand the problem") from tests alone. And if we are to believe Frederick Brooks, as I think we should, a good data model is imperative to good code. That is why TDD failed in the real world, and that problem is excarbated with agentic coding.
It is time we as an engineering discipline sober up about which methodologies work in the real world after decades of empirical data. Name the ten most successful software projects right in front of you now, and I bet that exactly none of them were designed from tests. One would expect at least one, from decades of best selling books on the subject.
In fact, it is high time to put the OOP paradigm under scrutiny and see which promises were delivered on and which weren't. Apart from a few functional languages, mainstream language design got stale for a long time until we finally got Rust and Go which broke with the strict paradigm.
TDD is still off though, clean room is better. With TDD your agent can cheat by making the implementation just pass tests without actually making the implementation do anything else otherwise. With clean room, you have agents develop implementation and tests independently, and have an arbiter run the tests, delivering feedback to test and implementation agents if they are wrong (yes, tests can be wrong also), no cheating, you need a specification as a single source of truth for this to work though.
I think you’re just lucky. Look at the Claude status page to see just how often they have outages (it’s almost daily). Even most of the green days have issues if you hover over them, they just don’t count them as outages.
Fable is better than Opus which is better than Sonnet which is better than Haiku. They’re basically just sizes.
Though it gets even more confusing because they also have effort levels so it’s not really possible to call one fast and one slow since Fable on Medium will be faster than Opus on Max.
I agree it’s confusing, and now OpenAI is following Anthropic’s lead with their new naming (Sol, Terra, Luna).
It's really not all that confusing. It takes 5 minutes to understand. Optimizing for absolutely no effort needed is silly. It's a thing, a topic, a skill, a domain. You have to get a little bit familiar with the terms in order to use it. Everything works like that. It's not that hard. The learning curve is very graceful. You can literally just start by asking any chatbot what the names mean. It's that easy.
A similar complaint was valid years ago when OpenAI had GPT-4o, o1, o3 (but no o2), o4-mini-high, GPT-4, and GPT-4.1 and GPT-3.5 etc.
Ok, the boring way would be a subset of XXS, XS, S, M, L, XL, XXL like clothes sizes. But it loses some marketing appeal and a quirky touch of personality that companies like.
Some models like ViTs use something similar but then introduce words with no unambiguous order, like Small, Medium/Base, Large but then I always forget if Huge or Giant is larger.
Maybe more accurately I should have said “larger”. Fable has the most parameters, Haiku has the fewest.
Also fwiw I’ve never found LLM benchmarks to match reality based on my own usage, not for the large frontier models or smaller open weight models so who knows if Opus is actually better than Fable (I doubt it).
Slowing down your internet connection due to using your bandwidth without your knowledge, getting blocked by websites due to malicious traffic from your IP address, extra annoying captchas due to the same. Those were just the first 3 things off the top of my head.
Got plenty of bandwidth, sites block you regardless, you get captchas regardless. Plus, IP addresses rotate once a day so you're just as likely to get someone else's bad reputation anyway.
To sue them you'd have to show some actual, concrete harm. For example you contact a site that is blocking you and they tell you it's due to a certain request and you trace that request to the proxy.
This is entirely specific to the user’s ISP, my “dynamic” IP goes months without changing.
And I would be miserable if I were subjected to the volume of captchas and blocks on my regular internet connection that I see on the occasion I use Tor.
Right? This is basically one small step away from stuff like “2TB usb stick for $20” scam listings on AliExpress. Of course it will be fake or crap. Cool tear down and write up though.
ZFS works fine on root on Linux. I use it on multiple machines.
> Linux is feeling more and more like a bunch of random tools thrown together as opposed to a complete OS designed to work as a whole.
This has quite literally always been the case and seems intentional. Linux is the kernel only, others supply the userland of their choice (aka a Linux distribution).
> Everything in the set was real. We couldn't fake any of it, because audiences are so sophisticated now in their knowledge of computers.
It’s funny he said that because when I first saw Jurassic park as a computer nerd kid, I was calling out the “this is UNIX, I know this” scene where she then flies around the file system in a 3d rendered file browser as typical movie computer BS.
Turns out it was a real application running on a real SGI machine, and ironically I was calling it out specifically because of my knowledge of computers (not having any idea about SGI machines at the time but having a ton of experience with DOS and Windows 3.1). My family who didn’t have much computer knowledge didn’t think anything of it.
I was going to reply that I thought the Boeing 737 was also hands down the most popular commercial airplane model in current use, so of course it would have more high profile issues. Then I did a quick web search and apparently the Airbus A320 is actually more popular since 2025, otherwise they’re fairly close to even. Yet these things only seem to happen to Boeing.
So yeah, never mind, fair game to blame Boeing actually.
I was never a fan of TDD, but ironically I think it may finally be a good development strategy now with agentic/LLM coding since LLMs do so much better they they have a verifiable goal like passing tests.