gk1

Exactly as happened with computer revolution... Expectations raised in line with productivity. In HN parlance, being a 10x engineer just becomes "being an engineer," and "100x engineer" is the new 10x engineer. And from what I can see in myself and others right now, being a 100x of anything, while exhilarating, is also mentally and physically taxing.

What are some striking examples from your experience?

BTW this is what I love most about HN - the surprising variety of people you can learn from, from billionaire founders to expat bingo-card geeks to Georgian-onion sellers to Dutch pro cleaners...

I actually use it all the time when driving a rental.

Love this announcement style. Direct, confident, and not a word longer than it needs to be. Gives major "the work speaks for itself" vibes. OpenAI's comms used to be like this, until it morphed into Apple-like grandiosity that instead comes off as try-hard.

@dang Any guidelines on obvious AI slop like this?

Exactly this. At least in the US, consultancies that contract with the gov’t can keep a small full time staff in order to qualify for small-business preference and keep their overhead low, and then depend on an army of subcontractors for large projects.

It’s not a mistake, it’s a choice. Quite a common one among younger people who grew up using Slack at work and iMessage everywhere else.

This comment is sickening. If this is coming from a personal hardship, I hope you find peace soon.

There are.[1] Unfortunately they take longer to employ than the crew had time.

[1] As it happens I open with an anecdote about steering redundancy on ships in this post: https://www.gkogan.co/simple-systems/

Aaaaand it's down again.

Founder of a/b testing company accuses founder of evals company of misrepresenting how a/b tests are used in practice, then concludes by misrepresenting how evals are used in practice: "Or you can write 10,000,000 evals."

Could've easily been framed as "you need both evals and a/b testing," but instead they chose this route which comes across as defensive, disingenuous, and desperate.

BTW, if a competitor ever writes a whole post to refute something you barely alluded to without even mentioning their name... congratulations, you've won.

You don’t know how good they were really doing and you don’t know how good the offer was.

From what I know they were the most used inference provider by developers a few years ago, but since the Together AI and Fireworks only grew while Replicate seems to have stayed quiet. It’s a highly competitive low-margin business, so volume is critical and if you’re losing volume then you’re doomed.

You know you have a branding problem when (1) you have to say that at the outset, and (2) it induces more eyerolls than a gaggle of golf dads.

The Augmentation vs Abdication framing is great. It also reflects my own experience in a different field: marketing. More specifically, coming up with messaging & positioning.

If you want something clear and compelling you can’t one-shot it, so you end up spending all day and night going back and forth with Gemini or Claude or whatever your cup of tea…

In the end I had something that looked great from the surface, but I felt completely lost in it. I couldn’t tell you what’s the core idea underpinning the whole thing, let alone explain it convincingly to the founders I work with.

The feeling is unsettling, but it has a benefit: it’s a signal that you’ve gone too far and need to re-calibrate.

In the end I found a great balance between the AI’s ability to synthesize vast information and offer endless alternatives to phrasing, with my personal taste and intuition. Augmentation.

They don’t need “many” people to fall for it. It’s a numbers game. Spam the message to 10k emails and even a small conversion rate can be profitable.

Also, I’d bet the average site owner does not know what a terminal is. Think small business owners. Plus the thought of losing revenue because their site is unusable injects a level of urgency which means they’re less likely to stop and think about what they’re doing.

Probably decided by committee and a subcommittee for each state. So you end up with inconsistent decisions between states.

They’ve had multiple secondary sales opportunities in the past few years, always at a higher valuation. By this point, if someone who’s been there >2 years hasn’t taken money off the table it’s most likely their decision.

I don’t work there but know several early folks and I’m absolutely thrilled for them.

R in RAG is for retrieval… of any kind. It doesn’t have to be vector search.

What do you propose they do? Because although something like Strata makes it easier, the reality is people are piling up MCP servers like they're free cupcakes. There's no getting the cat back in the box.

(I'm not in security so I genuinely don't know and am curious.)

Way less clear. Anthropic did it right and wrote about the “so what” instead of focusing on the underlying mechanics.

If the infamous Dropbox comment was a blog post in 2025.

(https://news.ycombinator.com/item?id=9224)

Seems successful to me… rolled it out to ~6% of their locations and got a ton of useful data, 2M successful orders (according to them), and kinda-funny viral marketing out of the small number of failed orders

Absolutely. Especially for late-stage candidates.

As someone who’s hired many dev advocates, I definitely value the ability to turn mundane topics into posts that hit the HN front page. If they can do this about something as dull as failing interviews, imagine what they’d do with an actually interesting technical topic.

What you’re describing is the result of the issue being complicated, not obvious.

Plus it’s straightforward to make Claude Code run agents in parallel/background just like Codex and Cursor, in local sandboxes: https://github.com/dagger/container-use

Not Windsurf… OpenAI. And OpenAI cares because they’re competing (in part) against Copilot, so if Msoft gets all the benefits of Windsurf then OpenAI would effectively be paying 3B to feed their competitor.