What the landscape means
The mountains on this site are real places, drawn by code from public elevation data under each place’s real sun, with no camera and no image model. They stand for a business seeing past how it runs today, ours included.
13 October 2026
Contents
Each mountain on this site has its coordinates in its caption. That’s the spot it’s drawn from, and you can find it on a map.
A script stands at that spot and faces the direction given in the table below. It reads public elevation survey data, the measured height of the ground at a fixed spacing, and draws one line across the picture for each step of distance away from you.
The texture comes from light. Each place is lit by its real sun, on the dates and at the hour the table gives. The script works out which slopes face the sun and which fall in another ridge’s shadow. Slopes facing the sun get heavier, brighter lines. Slopes facing away, or in shadow, get thinner, broken lines, or none.
What each picture is drawn from
Sun worked out with NOAA’s solar position formulas · Snowy Hydro, Snow depths · CSIRO, Climate concerns: Trends in Australian snow, 11 June 2024 · Ngāi Tahu Claims Settlement Act 1998, Schedule 75 · all read 11 October 2026
| Detail | The Hazards | The Remarkables | The Snowy Mountains |
|---|---|---|---|
| Where | Tasmania | Aotearoa New Zealand | New South Wales |
| Seen from | The water in Coles Bay | Lake Wakatipu, south of Kelvin Peninsula | A knoll 1,080 m up the Western Fall |
| Looking | Southeast | East-northeast | East-northeast |
| Depth | 7 km | 15 km | 12 km |
| Survey | 30 m grid | 8 m grid | 5 m grid |
| Heights | Stretched 1.55 times | True, narrow creek mouths filled in | Stretched 1.4 times |
| Water | Shoreline surveyed, swell modelled | Shoreline surveyed, ripple modelled | None drawn |
| Snow | None drawn | None drawn | Modelled from snow gauges |
| Date | 1 to 4 August, about 9:35 am | Mid-June to early July, about 3:10 pm | 14 June 2022, 2:30 pm |
| Sun | Northeast, 20° up | Northwest, 14° up | Northwest, 22° up |
The survey has no snow in it, so the Hazards and the Remarkables show bare ground. On the Snowy Mountains a model adds the snow, and it covers 40 to 46% of the picture, depending on the screen.
Snowy Hydro publishes snow depths for three sites, and CSIRO gives their heights as 1,460 to 1,840 m. In the week of 14 June 2022 the lowest site had the least snow. Carry that trend on downhill and the snow runs out a little above 1,100 m. So the model starts the snow at about 1,150 m, lower on shaded slopes and higher on sunny ones. Above the 1,800 m treeline it lies open, and below it the forest is drawn darker and broken. That’s why this picture is pinned to one real day while the other two cover a range of dates.
The open sea has no place behind it. It’s a wave model: a 12 second swell with two smaller swells across it and wind chop on top, each wave at its true length and speed, seen from 6 m above the water under a sun 6° up.
The Hazards stand on palawa Country, in Freycinet National Park. Lake Whakatipu-wai-Māori, under the Remarkables, carries a Ngāi Tahu statutory acknowledgement. The Western Fall is in Kosciuszko National Park, on Wiradjuri, Wolgalu, Ngunnawal and Monaro Ngarigo Country. The survey behind each place is credited at the end of this article.
Why lines
Geoscience Australia’s ELVIS, coverage map for elevation models · Spatial Services (DCS), NSW Elevation and Depth Theme · both read 11 October 2026
The first mountains we put out came from an image model. On 30 September a panorama of the Snowy Mountains near Perisher went up across three posts on our Instagram, with Instagram’s AI label on. The glacier lake that led our home page before this design came from the same kind of model.
In October we kept the natural look and asked for lakes, beaches and mountains drawn in lines.
We drew them from real ground. Every ridge sits where the survey data puts it, under a sun you can check against the dates in the table above. The set went through eleven rounds of changes, with fresh AI auditors scoring it between rounds on texture, line quality, motion and honesty.
The Snowy Mountains are back in lines, this time the Western Fall, on the other side of the Main Range from Perisher. They took three rounds of their own before two fresh AI auditors, one for the drawing and one for the code, each scored them 8 out of 10 or better.
The program that draws every line was written with Claude Code, Anthropic’s AI coding tool, the way we build software for clients. No AI model runs while a picture draws, so in our terms the pictures are built with AI.
As of October 2026 we can’t draw a beach that reads as one. A beach is a strip of sand 50 to 100 m wide that rises a few metres. We worked on one in Victoria from the third round to the eleventh, on 5 m survey data, and seen from the dunes, the sand still came out as one more band of lines next to the surf. Sydney’s beaches are 30 m data in the free tiles we use. New South Wales publishes free 1 m elevation data made from LiDAR, and on the national coverage map it covers all 21 Sydney ocean beaches we checked, Palm Beach to Cronulla. That’s what we’d try next.
Why mountains
We picked mountains on 29 September, while our pictures still came from an image model. They stood for reaching your potential, and for an owner seeing past the comfort of how the business runs today.
So every picture leads further in, toward a range or out over the water, with more past the edge of the frame. No one stands on top in any of them.
The same rules in your system
| In the pictures | In a system we build for you |
|---|---|
| Every ridge sits where the survey data puts it. | We test it on your real work before it goes live. |
| Each caption says when the heights are stretched or the snow is modelled. | We tell you which parts call an AI model as they work and which were only built with AI. |
| We can’t draw a beach yet, and the note above says so. | If a spreadsheet would fix it, or AI isn’t the answer, we’ll say so on the call. |
| The set was scored by fresh AI auditors that didn’t write the code. | A fresh AI auditor that didn’t build it checks it before you see it, and we fix what it finds. |
You can hold anyone who builds for you to that right-hand column, us included.
- Where does each part get its facts? A good answer names a source you can open, like your inbox or your job sheets.
- What has been stretched or filled in? With AI in the system, ask what happens when a customer leaves a detail out. A model can fill in a suburb they never wrote. You want the gap left empty and the customer flagged for you.
- What can’t it do yet? There’s always something. A builder who can’t name it hasn’t tested far enough.
- Who checked it, other than whoever built it? Whoever builds something checks for what they expect to find.
In your business, seeing past how it runs today starts with one job, the one that costs you the most hours every week. Get it fixed and working, then pick the next. The aim is a week with room in it to work on where the business is going. Where to start with AI in your business covers how to find that first job.
What they mean for us
Mountains are my place. I didn’t feel built for them, and they’re where I proved myself wrong. It’s the meaning that matters most here: going further than you thought you could.
The Snowy Mountains lead our About page. They’re the one range on this site in New South Wales, where Ostley is based.
Ostley is young, and we’re still learning how to scale a business, starting with this one. We run it on the same kind of system we build for clients, and each fix we make for one business teaches us something for the next. How we run Ostley shows that system.
The meaning has grown with the business. In September the mountains stood for reaching further. Drawing them from real ground in October gave them a second job. We don’t put a number on a page we can’t back, and now the pictures hold to that rule too.