Build1 distinct publisher3 min readPublished
Sending a six digit code is four jobs and only delivery reaches AWS. The deadline turns that one job into a choice between a shared sender identity you can use today and a leased number that pays off past a thousand messages a month.
The Engineer · Build desk

Compiled by The EngineerSomething wrong?How this is made
The cost table in that writeup is the part worth doing arithmetic on. It reports the two paths within a rounding error at 500 messages a month, and more than $100 a month apart by 3,000 [8]. Divide the gap by the volume and the managed path costs about 3.3 cents a message more [10]. Hold that delta at the low end and 500 messages would save roughly $16.70 of message cost, which is the amount the author says the fixed lease eats [8], so the long code lease in his model sits near seventeen dollars a month [11]. Parity at 500 is therefore the break-even, and every message above it is that lease being spread thinner [8].
For the shape to transfer to your account, several things have to hold. Every message goes to an Australian mobile number at standard pricing rather than a negotiated rate, and you are comparing a single long code lease [8]. The x-axis is messages, not logins. Resends and failed deliveries are messages, so a forecast built from expected sign-ins puts your crossover later than it will actually arrive.
The service menu is less tangled than the search results make it look. SNS Publish still sends, but since late 2024 its SMS traffic runs through AWS's newer messaging backend [3], so choosing SNS chooses a thinner API over the same delivery engine. Origination identities and configuration sets live in End User Messaging [4]. Notify sits on top of that same service with a shared, pre-approved sender identity and built-in fraud protection, and that shared identity is precisely what buys you out of the registration queue [6]. Your own number is cheaper per message and slower to start, because carrier registration takes real calendar time to clear [7]. There is also a send-and-verify pair where AWS generates and checks the code, billed per successful verification [5]; that is the one line item priced on users who complete a login rather than on attempts.
Then the defect. Creating a Notify configuration in the console returned ACCESS_DENIED_CONTACT_SUPPORT with no stated reason, while the identical configuration created through the CLI went active immediately, and the author says he has no confirmed explanation [9]. The account plainly held authority to create the thing, because it did [13]. The refusal traces to the console's path around the create call, separate from the create action itself. The console is an API client with opinions.
What the writeup does not carry is the rule. It gives the date and says the AWS side changes, without setting out what must be registered or by whom [12]. That schedules the decision but leaves the work unsized.
Ranked by verification strength, evidence, and original report placement.
Based on standard SMS pricing to Australian numbers plus a typical long code lease fee, the author reports that at 500 messages a month the two options are basically a rounding error apart because the fixed lease eats most of the saving, that past a thousand or so the gap opens up fast, and that by 3,000 messages the difference is over $100 a month.
A dev.to writeup describes building OTP delivery on AWS as Australia's new SMS Sender ID Register began reshaping delivery, and states that Australia's 1 July 2026 deadline changes the AWS side of sending a one-time code.
The author breaks OTP into four jobs: generate a short random code, store it tied to the session with an expiry usually of 5 to 10 minutes, deliver it almost always over SMS, and verify the typed input against storage and expiry. Generate, store and verify are described as plain application logic with nothing AWS specific about them; delivery is the interesting engineering decision.
Amazon SNS was historically the go-to for sending an SMS via its Publish API, but since late 2024 SNS SMS sending routes through AWS's newer messaging backend under the hood, making it more of a thin pub/sub wrapper than a separate delivery engine.
AWS End User Messaging (SMS) is the direct, current API for sending SMS, with full control over origination identities, configuration sets and two-way messaging.
AWS also offers an OTP-specific API pair, send and verify, in which AWS generates and checks the code for a fee per successful verification; the author says teams that already have generate and verify logic can skip it.
Distinct publishers with included, body-backed reporting in this cluster.
dev.to
1 article · September 2, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
build
Force the tool call, then hand Lightsail a long-lived key1 distinct publisher
build
CSA's 2026 threat list is a flat line, so ask which threats a config snapshot can prove1 distinct publisher
build
CSPM ends up as the intake queue for agent-provisioned infrastructure1 distinct publisher
build
Axonius runs one agent per tenant on AgentCore, and tracks model cost the same way1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One builder's notebook
Everything rests on a single dev.to post written in the first person, and it never cites the two documents that would settle it: AWS's pricing and service pages, and ACMA's own guidance. That is tolerable for the console error, which is testimony by nature, and thin for the rest — the SNS backend change 'since late 2024', the service hierarchy and the Australian rate assumptions are all checkable facts presented on trust. The text also breaks off mid-sentence at the most consequential point, where Notify's standing under the register was about to be explained.
A single account, twice
The only deployment anyone can point to is the author's own: one Notify configuration, refused by the console and accepted by the CLI. No customer count, no message volume actually sent, no second practitioner reporting the same console behaviour. Real usage of Notify against Australian numbers under the new regime is, on this reporting, unmeasured.
Slightly ahead of what it shows
The writeup is unusually well-hedged — 'run your own numbers', 'I don't have a confirmed explanation' — so the overstatement is structural rather than rhetorical. Two seams: a dollar figure quoted to the hundred with no rate card behind it, and a tighter coupling between 'lease your own number' and 'own a registered brand sender ID' than the piece ever demonstrates. The register governs the name at the top of the text, not the long code you rent, and the step from one to the other is exactly where the text stops.
Nothing being sold
A practitioner publishing what he learned on a developer blog, with no product, no affiliate path and no vendor disclosure in sight; he even tells readers to skip an AWS service they would otherwise pay for. The mild pull is reputational and platform-shaped: dev.to rewards a tidy narrative, and a post scoped to 'how to do this on AWS' has no reason to mention that Twilio and the local aggregators face the same 1 July 2026 rule.
Confident about the shape, not the digits
We can be fairly sure what this reporting says and reasonably sure of the parts that are public record and easy to verify — the deadline, the 'Unverified' label, which API is current. Confidence drops where it should: a single unreplicated console anomaly, and cost claims whose inputs were never shown. With no second publisher anywhere in this story, nothing here has survived a challenge.