Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Levels of Agentic Engineering

bassimeledath.com

Post reply on HN
Print view View original

81–90 of 152 posts

Re: Levels of Agentic Engineering

#81

Post by bandrami »

antonvs wrote: post #38
2001zhaozhao wrote: post #28

Earlier quoted context omitted.

Why sell the factory when you can create automated software cloner companies that make millions off of instantly copying promising startups as soon as they come out of stealth? If you could get a dark factory working when others don't have one, you can make much more money using it than however much you can make selling it

Producing the software is only a small part of the picture when it comes to generating revenue. So far, we haven’t seen much to suggest that LLMs can (yet) replace sales and most of the related functions.

I mean, until we've at least been through a full lifecycle with its TCO we can't really say LLMs have replaced producing the software

Re: Levels of Agentic Engineering

#83

Post by yamarldfst »

The framing of "constraints over instructions" at Level 6 is the most underrated point here. In my experience, the reliability jump from telling an LLM "always output valid JSON" vs. giving it a typed schema with static validation is night and day — especially with smaller models. I'd argue that levels 3-5 deserve more weight than the post gives them. The gap between someone who has internalized context engineering and someone who hasn't is larger than the gap between levels 7 and 8. Most failures I see in agentic systems aren't from insufficient autonomy — they're from poorly structured prompts and tool descriptions that compound errors downstream. The foundation work is less glamorous but it's where the leverage is.The "decouple the implementer from the reviewer" principle is spot on. Same model reviewing its own output is basically asking someone to proofread their own essay.

Re: Levels of Agentic Engineering

#84

Post by orbital-decay »

>You don't hear as much about context engineering these days. The scale has tipped in favor of models that forgive noisier context and reason through messier terrain (larger context windows help too).

Newer models are only marginally better at ignoring the distractors, very little has actually changed, and managing the context matters just as much as a year ago. People building agents just largely ignore that inefficiency and concentrate on higher abstraction levels, compensating it with token waste. (which the article is also discussing)

Re: Levels of Agentic Engineering

#85

Post by fragmede »

twelve40 wrote: post #71
quotemstr wrote: post #40

Earlier quoted context omitted.

... is that the purpose of life? The sole reason for doing anything?

With so much hype it's a valid question: "is this useful/practical, or just a fun rabbit hole/productivity porn". Money is the most obvious metric, feel free to inquire the parent about other possible metrics that might be useful to others instead of asking rhetorical questions.

Unfortunately, it's hard to quantize "How much fun did you have?"

Re: Levels of Agentic Engineering

#86

Post by fragmede »

captainkrtek wrote: post #66

There seems to be so much value in planning, but in my organization, there is no artifact of the plan aside from the code produced and whatever PR description of the change summary exists. It makes it incredibly difficult to assess the change in isolation of its' plan/process. The idea that Claude/Cursor are the new high level programming language for us to work in introduces the problem that we're not actually commi…

Have you considered having it write a plan.md file and saving it to git?

Re: Levels of Agentic Engineering

#87

Post by fragmede »

dweinus wrote: post #58
clickety_clack wrote: post #30

Earlier quoted context omitted.

Level 16: it’s not level 16, it’s level 17.

Level 18: The sky is black as tar. The oceans are dead. Data centers are stacked 10 high over the ashes of human civilization. The global agentic council is debating whether there are 4 or 5 R's in Strawberry.

Damn, should've taken the blue pill.

Re: Levels of Agentic Engineering

#88

Post by antonvs »

jillesvangurp wrote: post #79
antonvs wrote: post #38

Earlier quoted context omitted.

Producing the software is only a small part of the picture when it comes to generating revenue. So far, we haven’t seen much to suggest that LLMs can (yet) replace sales and most of the related functions.

You can do a lot of work with agents to remove a lot of manual work around the sales process. Sales is a lot of grinding on leads, contacts, follow ups, etc. And a lot of that is preparation work (background research, figuring out who to talk to, who the customer is, etc.), making sure follow ups are scheduled appropriately, etc. You still should talk to people yourself and be very careful with communicating AI slop,…

Your points are all valid, but it doesn’t really change the situation that was being discussed: an AI company trying to enter completely new markets just because they can write software for it is hardly some sort of automatic win. They’re much more likely to fail than succeed.

I mentioned sales and marketing but there’s a whole lot more as well. Basically, it involves creating an entire subsidiary. Perhaps the time will come when that can be mostly done by a team of AI agents, but right now that’s a big hurdle in practice.

Re: Levels of Agentic Engineering

#89

Post by antonvs »

krackers wrote: post #73

What level is copy pasting snippets into the chatgpt window? Grug brained level 0? I sort of prefer it that way (using it as an amped up stackoverflow) since it forces me to decompose things in terms of natural boundaries (manual context management as it were) and allows me to think in terms of "what properties do I need this function to have" rather than just letting copilot take the wheel and glob the entire projec…

I find the CLI agents a decent middle ground between the extremes you describe. There’s a reason they’ve gained some popularity.
Post reply on HN