ZephyrBlu

7 ms·

1.
by ZephyrBlu 11d ago
Thinking of AI-generated code and AI code reviews as the same "AI" is not correct. The reviewer is using a fresh context window with no previous knowledge of the changes. That is why it's powerful, because you get the agent t
2.
by ZephyrBlu 4mo ago
I completely agree with your distinction and that is exactly what they mandated :) I don't think that is what most people colloquially mean by "feature flags" though. Even most teams in Shopify abused "ephemeral" fl
3.
by ZephyrBlu 4mo ago
In a company the size of Shopify people flipping their feature flags would very often impact *other teams*, and like I said feature flags got abused with even seemingly innocuous changes being put behind them or being left long periods of t
4.
by ZephyrBlu 4mo ago
Feature flags need to be treated as short-lived and experimental otherwise they end up getting abused for everything and make it very difficult to reason about your application. If it's config/customization, it should be in code.
5.
by ZephyrBlu 4mo ago
No other provider works like Copilot did with "premium requests". Usage limits (Codex/Claude Code), which are inherently linked to tokens, are the most common. Some providers like Amp charge you per-token like Copilot is movi
6.
by ZephyrBlu 4mo ago
Why would you extrapolate from Microsoft's very poor setup to tokens in general then if you know it's stupid and not representative?
7.
by ZephyrBlu 4mo ago
The fact you are trying to use Copilot as an example here shows you don't understand how Copilot's previous billing worked. Previously they used "premium requests" which would allow you to make a request to one of the mo
8.
by ZephyrBlu 5mo ago
Scheduled maintenance in 2026 is insane
9.
by ZephyrBlu 5mo ago
No point in discussing with someone who is arguing in bad faith. I already agreed that some parts of the engineering process are safety critical. If you think there is no bullshit in the process you don't have enough knowledge about th
10.
by ZephyrBlu 5mo ago
People like to make this point, but traditional engineering has the opposite problem: insanely overwrought processes and box-checking that exists for no reason and slows everything down to a snail's pace. Yes there are safety-critical
11.
by ZephyrBlu 8mo ago
No it was the 3 co-founders, a part-time person and 4 engineers. Now they are 3 engineers down.
12.
by ZephyrBlu 8mo ago
He said he wanted to state it like that because he thought just saying "3 people" undersold the impact.
13.
by ZephyrBlu 8mo ago
On his morning walk/podcast thing about the topic he said 75% of the team = 3 developers
14.
by ZephyrBlu 8mo ago
Definitely more than 200k per head. I remember seeing a job posting for Tailwind Labs for a (design?) engineer which was 250-300k TC. Seems like it was an insanely profitable product, but a risky business.
15.
by ZephyrBlu 10mo ago
Material and cut/design. Material is not just about quality, but rarity or uniqueness. For example, japanese denim can get very expensive in part because it's very low volume. For dress pants, it might be a particularly interestin
16.
by ZephyrBlu 11mo ago
Because it's not his job. He should elevate someone else into that IC role instead of holding it for himself. The way he describes it, there is no one else in the company who can do the IC work he is doing, which is long-term bad. Codi
17.
by ZephyrBlu 11mo ago
I wouldn't really call it "demand". It's more like one-shotting humans with a product which maximally stimulates them through what is basically a psychological hack. We were not built with the capacity to handle the shee
18.
by ZephyrBlu 11mo ago
The thing I'm most curious about from this article is how/why the author was demoted from E9 to E7. A demotion in itself is pretty unusual, but being bumped down 2 levels seems super weird. E: ok watched an interview the author ga
19.
by ZephyrBlu 11mo ago
I used both English and Hangul to search. Searching for general things like food was good, but if I was trying to find a specific address it was very difficult. Sometimes it would just return completely wrong garbage. One time I was trying
20.
by ZephyrBlu 1y ago
They lack a lot of polish. Functionally they're mostly usable, but some interactions are janky and I found the search to be super hit or miss.
21.
by ZephyrBlu 1y ago
Pretty crazy improvement, seems like he's putting his money where his mouth is
22.
by ZephyrBlu 1y ago
On vibe coding being self-correcting, I would point to the growing number of companies mandating usage of AI and the quote "the market can stay irrational longer than you can stay solvent". Companies routinely burn millions of dol
23.
by ZephyrBlu 1y ago
> No it isn't. There's literally nothing about the process that forces you to skip understanding. Any such skips are purely due to the lack of will on the developer's side This is the whole point. The marginal dev will g
24.
by ZephyrBlu 1y ago
From the title and URL I thought the substack was "Big Tech Public Relations"
25.
by ZephyrBlu 1y ago
__350x-ing__, not 35
26.
by ZephyrBlu 1y ago
I tend to agree, which makes it all the more amusing that companies brag about being so selective. It seems like largely artificial and random selectivity.
27.
by ZephyrBlu 1y ago
I'm pretty sure it's because an image of a banana under a microscope generated by the model went super viral
28.
by ZephyrBlu 1y ago
I'm confused by this post because I think Sorbet satisfies basically all the things the author wants, and my experience with Sorbet has been really good!
29.
by ZephyrBlu 1y ago
He is not a "courseboi". SaaStr is a legit brand that's been around for a long time focusing on the sales side of SaaS. You have to remember this is someone who is almost certainly completely non-technical and purely vibe cod
30.
by ZephyrBlu 1y ago
Smartphones as a lower bound is crazy

More ›