frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

You only need the frontier model for one single edit

https://stencil.so/blog/prewalk
120•jxmorris12•6d ago

Comments

pistoriusp•10h ago
Apparently "/prewalk" is built into https://github.com/can1357/oh-my-pi, which I guess is a lot like "Oh My ZSH?" Pi with batteries included.

There are a lot of interesting ideas in it, mostly none that I have applied to my own workflows. Curious if anyone has used it?

AshamedBadger56•9h ago
I've played around with all the popular agent CLI's and I've landed on omp/ohmypi as my favorite for now. It's definitely the opposite of a lean install like Pi, but I find it works a lot better than OpenCode. It's also very very active, with a lot of cool new ideas like this prewalk in it.
the_gipsy•9h ago
> Senior architect, junior engineer. Sounds great, right?

Sounds like a consulting slaughterhouse. Picture Java Enterprise Solutions. As we all know, the pinnacle of software engineering.

swiftcoder•1h ago
Hah. Yeah. Weirdly, most software managers also seem to believe this whole "senior writes a plan, junior implements" is cost-effective (it typically isn't, one should invest in hiring/promoting more seniors).
wrs•9h ago
Lately I'm trying a variation on this: Have the main agent make a phased implementation plan, then for each phase, have it start an implementation subagent with a focused prompt, then review that agent's work in the main session. The theory being that the main session still contains all the research, but it can review just the diff rather than have the entire implementation session in context as well.

The post doesn't include a review in the cost comparison, but I find that immediately doing a review catches lots of mistakes.

pqdbr•7h ago
I’m using Claude code dynamic workflow like this. I tell Fable to use a workflow. He is the planner, orchestrator and reviewer. Opus agents are implementers. Works unbelievably well.
danielciocirlan•5h ago
“He”
llbbdd•5h ago
I call him "my son"
oliver236•56m ago
can you explain this in a bit more detail? super interested
sgc•5h ago
Actually it does include an important level of review. They say "init a TODO list with a validation step for each item", and then the last several sections of the execution are within the TODO - where it is validating the code.

To me that was the heart of the article and of what they did - create a self-validating todo list that deterministically forces the agents to stay on task, and then trim out just the right extraneous context (the actual planning) without messing with the read-file context which is where they found subagents burning most of their tokens.

Just a few days ago I was speculating about forking context to implement instead of spawning subagents here on hn. This is similar to that just several more steps more sophisticated. You could even fork the trimmed context here for very large tasks.

Almondsetat•8h ago
Did they also use Gemini Flash to write this article? Because, frankly, it's unbearable
IshKebab•14m ago
Very clearly. Even if you have no qualms with AI I don't understand why people think writing in this cliched style is ok.

On the other hand people thought it was fine to use cliched titles like "The unreasonable effectiveness of..", "... for fun and profit", "The rise and fall of ...", "All you need is ...", etc. etc.

So I guess there's just a large portion of writers that are immune to cliche? Pretty annoying anyway.

rwc•8h ago
Just specify that the more efficient model should only be used if the cost of doing the task would be less than using the higher horsepower model. Fable has no problem routing based on that instruction.
viccis•7h ago
From what I can tell, I have a similar workflow. I get Fable 5 / Sol on High to talk to me about requirements until it's ready to design. I then have it break the design into tickets. I use kata, an agent-oriented issue tracker. If it ever starts to get bloated, I'll just make my own as it doesn't need too many features. Anyway, I tell it to include sufficient context in each such ticket to be picked up by a new implementation agent. Explicit user stories, Cucumber style acceptance criteria, etc.

When it's done, I switch to a smaller model and tell it to start a /goal of calling `kata ready` to get tickets ready to be worked. Work one ticket on each goal iteration, committing changes when done. Stop when all the tickets are either closed out or are blocked on actions from me.

It works fantastically. I can get entire (relatively straightforward) iOS apps done in under my $20/month five hour session window.

I get even better results if I do the QA session with the frontier model with two output artifacts: an implementation spec and a design prompt for Claude Design. I push the design prompt through Claude Design, tweak the results, and get a design spec. When I have the frontier model do the planning, I have it read the implementation spec from before, along with the design spec's overview file.

When I do it this way, I've done A/B tests between Fable 5 and Sol, and the apps they wrote were basically identical. It's so much cheaper than the "let loose the subagent fleet!" form of context management.

I haven't even added any kind of frontier model validation cycle into this loop yet. Everyone keeps talking about a subagent flow in which the big boy reviews the work of the drones, but I've found that if it encodes its acceptance criteria well enough, they do a satisfactory job of it themselves.

hankbond•7h ago
Does anyone have a name for that really neat "the bar chart is a bookshelf, you can hover a spine to see the cover" vis? I've never seen it before.
Bolwin•7h ago
This is a lot of words to say "switch models instead of writing a plan file and starting a new session" which I do anyway.

That said, it takes me a while to reads plans and often I'll take a break, by when the cache has expired. At that point it may be better to start over

fastball•7h ago
I don't think that is the takeaway at all.
pphysch•7h ago
> Any agent, any model, any scaffold: the bill is essentially O(reads)

Therefore token sellers have incentive to produce verbose, unreadable code

pixl97•6h ago
Whichever token seller defects can capture a larger chunk of the market with slimmer code.
pphysch•3h ago
Do you think the average vibe coder or AI-smitten suit cares about slim code? A lot of folks still think "LOC = good" in 2026. It will be a hard sell for the practical engineer.

Even if they understand "slim code = fewer tokens = less spend" they still have to grok that less code is not less functionality.

figmert•7h ago
The reason I use plan then implement is because I can adjust the plan, whereas if I get it to implement straight away, it might (and often does) make the wrong decisions that will be harder to adjust, or I'd have to adjust it after the fact.
canpan•6h ago
One big pain point is not seeing the thinking trace. I noticed using pi with a self hosted model that I could spot, stop and correct my prompt much faster. With claude etc I have to first wait for it to think 3 minutes and then notice it got completely off the path I wanted.

I still use a plan, because my local model is not as smart as opus, but I can iterate much faster.

carterschonwald•5h ago
this so true. its really hard to make sure a model isnt going off the rails if i dont see full cot. the fact that oai and anthropic models hide it now has made them less reliable. which is a shame
trollbridge•4h ago
The great news is you can use DS-V-Pro, MiMo-V2.5-Pro, GLM-5.2, or K3 and see everything.

I tend to use 5.6-Sol now more for one shot type of tasks where all I want is the answer and I’m not going to read the reasoning.

nchmy•7h ago
This entire argument rests upon the fact that Gemini Flash ain't cheap.

Try plan with opus 4.8 and then implement with Deepseek v4 flash - its 35x cheaper for reads, 90x cheaper writes, and 18x cheaper cache reads.

Or plan with Deepseek Pro, or even Flash itself. I've been impressed with both.

pdyc•1h ago
+1. i have similar workflow and i use local models on igpu so token cost is free and electricity cost is in cents.
ricardobeat•3h ago
I just did this five minutes ago. Started a task with Kimi K3, noticed cost going up, switched to Minimax M3 continuing from the same context. Really easy to do with Crush/OpenCode.

It works quite well most of the time. For very deeply technical tasks, where exploration involves multiple agents (and blowing up context limits a few times), I'll have a plan written down to disk that will be much longer than 2k tokens.

potus_kushner•2h ago
this article is full of missing text like this: " of Opus at of the cost, the speed, points over oneshot Flash." is it just me or is that page using some brand new js feature or chromium-only hack?
Too•1h ago
> The mistake is upstream of the architecture diagram. People price agents the way they price people: senior time is expensive, so minimize senior involvement.

> But the expensive part of an agent's day is not the fixing, building, or even the thinking. Opus fixing things does not cost money. Opus reading things costs money.

Heh, another rediscovery that agents and humans are alike. By the time I've researched a Jira ticket and made it unambiguous enough to outsource, I might as well have written the code myself instead.

wordpad•1h ago
You are more proficient at writing code than ai specs.

You could get better/faster at writing good specs and its a higher cap skill now.

For low risk changes you could let llm write its own spec from high level requirements and just validate its assumptions/design decisions.

egamirorrim•1h ago
Isn't part of this the inevitable cache invalidation that comes from switching between providers (Opus to Gemini)?
Aeolun•44m ago
Yeah, but you only pay the switch on the much cheaper model, the first bit is done with smaller context on the expensive model, so you still save money in the end.
nxtfari•1h ago
This is really smart, like the author said, old idea but cleverly applied.

In case anyone wants a summary: don’t one shot, don’t use plan mode and hand off the plan to cheap executors, ask the frontier model to explore, create a todo list, and then start when it feels confident; stop it after first code edit, then prefill the context to cheap executor to continue.

_ink_•36m ago
> Below: the share of runs that went poking around the web for it. Filthy cheaters!

I can understand that in their benchmark setting they wouldn't want the model to find an existing solution. But in my day to day work, wouldn't I want the model to search online for the best solution? Am I not shooting myself in the foot by preventing it?

andreyvit•8m ago
It's hard to argue with the numbers, but starting with a (true!) “research is the most expensive part” premise, this strikes me as an odd direction to go to optimize costs:

1. As others pointed out, we feed all the same research turns to a smaller model, so we pay the uncached price for all of them.

2. During research, the model typically reads more code than is relevant, to figure out what is relevant and what is not. If that's the expensive part, we keep paying for those turns with the most expensive model?

3. There is no quality comparison of the resulting code. Same plan != same code, and AI tokens during the initial implementation phase isn't the only cost attributable to the task.

I do the opposite:

1. Outsource research to a subagent, or several parallel subagents. Let them output the relevant code paths only. Using gpt-5.6-terra-high on Codex and sonnet on Claude. Merge results into a single per-task research file. This saves tokens and context window of the bigger model, and avoids re-researching after compactions and in subagents.

2. Use the smartest agent (Sol xhigh ultra / Fable max) for both planning and execution. Tell it to use the research file where possible.

3. Switch to dumber agents for verbose substeps, e.g. gpt-5.6-luna-medium / sonnet is enough to drive browser use.

Sadly, got no numbers to back this approach.

bensyverson•4h ago
Yes, this is a core part of my workflow; use the smartest model (e.g. Fable) to generate a large hierarchical implementation plan, then clear the context and have a lesser model (e.g. Sonnet) track and execute the plan [0], often in parallel.

It all depends on how much work you're doing; if it's a simple task, there's virtually no need for plans at all; just have the smart agent do it. But if the work is going to take 5, 10 or 100 sessions, there's real value in the "smart agent plans, cheap agent executes" model.

[0]: https://github.com/bensyverson/jobs

ricardobeat•3h ago
This is exactly what the post argues against though, as it leads to higher overall cost.

Incremental – A library for incremental computations

https://github.com/janestreet/incremental
133•handfuloflight•3h ago•19 comments

Who's afraid of Chinese models?

https://stratechery.com/2026/whos-afraid-of-chinese-models/
559•mfiguiere•20h ago•379 comments

Running Doom on Our Custom CPU and Going Viral

https://www.armaangomes.com/blogs/doom/
44•arghunter•3h ago•7 comments

A Koi Pond Mosaic Made from 10 Pounds of 3D Printer Waste

https://www.instructables.com/A-Koi-Pond-Mosaic-Made-From-10-Pounds-of-3D-Printe/
26•sudo_cowsay•3h ago•19 comments

Kimi Work

https://www.kimi.com/products/kimi-work
511•ms7892•14h ago•216 comments

Five US tech giants' hidden debts soar to $1.65T on opaque AI funding

https://asia.nikkei.com/business/technology/five-us-tech-giants-hidden-debts-soar-to-1.65tn-on-op...
211•NordStreamYacht•3h ago•81 comments

Jelly UI: Soft-body physics for native HTML form controls

https://jelly-ui.com/
451•baldvinmar•14h ago•147 comments

Human mathematicians are being outcounterexampled

https://xenaproject.wordpress.com/2026/07/20/human-mathematicians-are-being-outcounterexampled/
313•artninja1988•12h ago•122 comments

Show HN: Ex Situ – Open-source spatial index of displaced cultural artifacts

https://exsitu.app/map
22•hbyel•2h ago•6 comments

Nativ: Run frontier open models locally on your Mac

https://blaizzy.github.io/nativ/
256•aratahikaru5•13h ago•86 comments

Flock Credibility Lost as It Repeatedly Lies to City Councils, Police, & Public

https://www.aclu.org/news/privacy-technology/tracking-alpr-cameras/flock-safety-credibility-lost-...
352•StatsAreFun•7h ago•82 comments

Show HN: Immersive Gaussian Splat tour of grace cathedral, San Francisco

https://vincentwoo.com/3d/grace_cathedral/
147•akanet•11h ago•31 comments

Tensor Notation in Mathup

https://mathup.xyz/#tensors
7•runarberg•6d ago•1 comments

Linux kernel will support $ORIGIN, sort of

https://fzakaria.com/2026/07/20/linux-kernel-will-support-origin-sort-of
4•ingve•1h ago•1 comments

Agent swarms and the new model economics

https://cursor.com/blog/agent-swarm-model-economics
179•jlaneve•13h ago•82 comments

I wrote an bash enumerator because I was sick of xargs

https://numerlab.org/2025/07/20/bashumerate-enumerator/
118•wallach-game•11h ago•92 comments

Launch HN: Bloomy (YC S26) – AI-powered mastery learning for K-12

83•alexsouthmayd•15h ago•85 comments

China’s open-weights AI strategy is winning

https://werd.io/american-ai-is-locked-down-and-proprietary-its-losing/
1076•benwerd•17h ago•827 comments

Jellyfin founder Andrew leaves team

https://forum.jellyfin.org/t-project-leadership-changes
222•swat535•8h ago•156 comments

The Psychology of Software Teams

https://www.routledge.com/The-Psychology-of-Software-Teams/Hicks/p/book/9781032963389
82•dcre•5d ago•20 comments

You only need the frontier model for one single edit

https://stencil.so/blog/prewalk
120•jxmorris12•6d ago•37 comments

Flight Planning with Little Navmap

https://tech.marksblogg.com/little-navmap-flight-planning.html
11•marklit•4d ago•4 comments

The Power of Awareness: Overcoming Surveillance Capitalism

https://www.scottrlarson.com/presentations/overcoming-surveillance-capitalism-with-awareness/
97•trinsic2•11h ago•15 comments

My two year old taught me constraint solving

https://thecomputersciencebook.com/posts/how-my-2yo-taught-me-constraint-solving/
66•bambataa•1w ago•26 comments

Shinjuku Station in 3D

https://satoshi7190.github.io/Shinjuku-indoor-threejs-demo/
206•Gecko4072•17h ago•43 comments

Corners Don't Look Like That: Regarding Screenspace Ambient Occlusion (2012)

https://nothings.org/gamedev/ssao/
163•firephox•16h ago•68 comments

Hacker wipes Romania's land registry database

https://news.risky.biz/risky-bulletin-hacker-wipes-romanias-entire-land-registry-database/
634•speckx•18h ago•347 comments

How we measured AI writing across arXiv, and where the measurement breaks

https://unslop.run/blog/measuring-ai-writing-on-arxiv
215•dopamine_daddy•14h ago•152 comments

Perfection is not over-engineering

https://var0.xyz/posts/perfection-is-not-over-engineering.html
230•var0xyz•17h ago•99 comments

Claude Fable produced a counterexample to the Jacobian Conjecture

https://xcancel.com/__alpoge__/status/2079028340955197566
733•loubbrad•1d ago•468 comments