waynenilsen

> bot detectors made it extremely difficult

i am looking forward to when we can put this behind us, it is still a major issue

Cool I made this thing a while back but I really like your fork spawn parallelism

https://github.com/waynenilsen/crumbler

This uses recursive task decomposition but is single thread by design. Honestly fast enough for me and makes it easier to reason about

ASIC inference is clearly the future just as ASIC bitcoin mining was

i believe agent native sites will stand up and the incumbents will be forced to adapt

such as agent native shopping platforms whereby a human will bring you something from walmart or what not could spring up and disrupt your instacart of the world

this of course is just one simple example, when it works better for the clawdbot or whatever comes next what are the users going to choose they'll say 'get me some apples from walmart using instacartforbots' because they know the agent success rate will be higher

ah yes thank you very much that did work

still not great in my opinion sadly i think this will likely get worse and perhaps never get better

commit messages as marketing materials and 'kpi's is an anti-pattern i would have preferred to avoid

i can see a new token efficient mirror web possibly emerging using content type headers on the request side

forms, PRG, semantic HTML and no js needed

Frontend QA is the final frontier, good luck, you are over the target.

The amount of manual QA I am currently subjected to is simultaneously infuriating and hilarious. The foundation models are up to the task but we need new abstractions and layers to correctly fix it. This will all go the way of the dodo in 12 months but it'll be useful in the meantime.

agent-browser helped a lot over playwright but doesn't completely close the gap.

This post from 2022 is the one I keep going back to

https://david.rothlis.net/declarative-schema-migration-for-s...

> Information leakage may also be occurring via PDF comments or orphaned objects inside compressed object streams, as I discovered above.

hopefully someone is independently archiving all documents

my understanding is that some are being removed

i think when someone makes the cli like this they're going to win

$ phone call bill

ok call_id=3f2a

$ phone status 3f2a

dialing

$ phone status 3f2a

answered

bill: hello

$ phone say 3f2a "hey, quick question"

ok

amazing now do the same for voice and sms!

i see what you are saying now, i have actually started doing my own dockerfiles so everything is in one container it works reasonably well actually eg you install postgres redis x y and z on the same thing the agent is running on then you can start fresh anytime then all of your stuff is on localhost

> intent drifted

Indeed I believe this is probably the next thing to solve but even here I don't think it is out of reach. What we aught to be able to do is disconnect and make asynchronous the goals of the project with where we are. This, in normal software building, is encapsulated by the roadmap. I am building roadmapping prompts now and broadening the scope of the software development lifecycle even further to the encapsulate the roadmap as well which was previously out of scope for the experiment I am running now.

The prompts I am using now give the agent autonomy over 'make the next prd that makes sense' however I think it is a straightforward extension to add 'in the context of the @roadmap/ ' or similar with probably decent results.

Have you tried something similar?

Even without a roadmap the agent continues to do useful work over 24 hours in. You can see the commits and PRDs they really are quite sensible and I pulled and tested and everything really is working quite well. Frankly, I am shocked it is working at all. I have had to step in once or twice you definitely need to keep an eye on the logs every once in a while. Getting the loop booted up in a reliable way was the hardest part to be honest and even that was not terribly difficult.

https://github.com/waynenilsen/ralph-kata-2

say your docker compose file maps postgres to 55555 on your laptop then in the sandbox db url host is host.docker.internal and port is 55555

does that make sense?

i have seen the same, however, it can often easily find its own bugs when prompted to do so, in this case, with a ticket perhaps

the ticket burndown is a very nice feature because whenever you want to add a ticket it'll just pick it up and do its best

You can eject to host.docker.internal it’s the easiest way

> the LLMs aren't as good at making 3D objects as they are are writing code

I am still hoping that openSCAD or something similar can grab hold of the community. openSCAD needs some kind of npm as well as imports for mcmaster-carr etc but I think it could work.

Once you get your setup bulletproof such that you can have multiple agents running at the same time that can run unit tests and close their own loops things get even faster. However you accomplish that. Not as easy as it sounds mostly (and absurdly) due to port collision. E2E testing with playwright is another leap.

What kills me is I can’t double click the thing to select it.

This deeply resonates. I got out of the game. The issue is that financial freedom was never a technology problem. Meatspace will always get their hands on anything with a moderate amount of traction.

marketing.

Now do the same for ffi

comparisons with internet age very much resonate - dark compute will be as dark fiber was

The software requirements phase is becoming increasingly critical to the development lifecycle and that trend will continue. I have started writing very short tickets and having claude code inflate them, then I polish those. I often include negative prompts at this point so claude may have included "add a progress bar for xyz" and i simply add "do not" in front of those things that do not make sense. The results have been excellent.

i think its probably mostly vibes but that still counts, this is not in the charts

> Windsurf reports Opus 4.1 delivers a one standard deviation improvement over Opus 4 on their junior developer benchmark, showing roughly the same performance leap as the jump from Sonnet 3.7 to Sonnet 4.

also significantly better for performance

i need tree conversations more now than ever

I am very much looking forward to seeing how things evolve. Do you think the bitter lesson applies here? I am assuming you are gathering reams of data from the controls and any cameras of manually operated machines. I have been thinking about the bitter lesson in many contexts recently and I think for something like this where there really isn't anything proprietary about the method of operation, doing some kind of data consignment agreement or just up front paying for data could be very valuable in and of itself. Even unlabeled data can later be autolabeled.

do you see a change in the machinery landscape due to the removal of humans?

Review and good rules are still critical. Current agent state is still hyperspeed junior engineer. Providing examples helps a lot when scaffolding something similar to something else.