albert_e

Looks useful. I was experimenting with making an SVG creator tool that can churn out shapes that go beyond the shape library of microsoft powerpoint -- so my slides can stand out. Things like a hexagon with rounded corners, squircle, swooshes and blobs etc.

Patterns like these are useful as well.

The auto mode classifier has been out of service more than once in last 24 hours.

Had to shift+tab into other modes and manually approve permissions like a cave man

Security by obscurity -- such an age old anti-pattern!

I believe many AI tools like Gemini generate publicly accessible URLs when we click "Share" on any chat conversation -- and expect users to then own the lifecycle of that link

Depending on how the link gets handled -- by the browser, device OS, any hooks/plugins/extensions, aggressive telemetry, social media url previews, preload/prefetch, wrapping and url shortening, etc as it reaches the intended user -- there are countless ways in which the URL can be indexed and scraped

There was a issue not long ago when Claude artifacts were indexed en-masse by Google and other search engines

This is shockingly lax approach to data security and privacy by design

> If you accidentally send a request with say, ~700k context already accumulated in a session which is outside cache TTL, you're paying full cost of these 700k tokens.

Thinking aloud:

The harness UI should probably implement a timer that shows whether you are still within Cache TTL since your last turn of the conversation.

> What this is normally called

Is this also what Microsoft calls "Magentic" pattern?

(for a long time i kept reading it as Magnetic pattern)

https://learn.microsoft.com/en-us/semantic-kernel/frameworks...

My hypothesis:

One way to bias the generative output away from ugly generic sloppy defaults and towards something you personally prefer and desire based on your taste ... is to show it examples of what good looks like to you.

It is very hard for us to describe in words why you like a certain thing. But you can spot it when you see it. Or you can describe some aspects of it that you like or dislike after you see it -- but not before.

This approach is more accessible to casual creators -- who do not have formal training and experience to say "use corporate memphis" or "punk fanzine look with grunge typography".

(and this can apply to many fields -- not just visual design)

If this continues to be a thing ...one might actually imagine a good system where we have AGENTS.md with all generic agent instructions and a bunch of harness specific files like CLAUDE.md, GEMINI.md, and CODEX.md that all add whatever nuance and nudging each harness seems to individually require

Should we be able to manually rename those history folders to make the session history show up as --resume options in a new folder?

Fair enough.

So the chain of trust is ...

Personal credibility of the person taking the photo with iPhone (till the moment picture is taken) + tamper-evident protection against unauthorized or anonymous edits and manipulation (after the iphone photo is published and circulated)

Apple cannot (of course) vouch for the former.

AWS Fault Injection Service

https://aws.amazon.com/fis/

https://docs.aws.amazon.com/fis/latest/userguide/what-is.htm...

But alongside the marketing blitz they also offer certifications for people who are supposed to execute these projects. Even the most foundational certification exam -- which simply tests your recall on what AWS service is for compute versus networking -- teaches and tests you about "Shared Responsibility Model" that draws the line at what the customer is still responsible for when they use cloud bases IaaS / PaaS / SaaS services.

If businesses are going to cloud but without engaging / listening to competent people who know these basics -- then the blame needs to be somewhat pointed back at those very business leaders I feel.

This is not obscure magical knowledge either that is tightly controlled. Any cloud vendor will freely teach you that. Or even a google search would.

Thinking aloud about failure modes / edge cases:

If i create a high quality deepfake image, project it on a large screen and take a photo of that image with an iPhone (+apple verified image secure tag) ... would that resulting image be considered "authentic" by default?

Would digital forensics accessible to lay people still be able to fact-check and call out misuse / fakery of the new secure tag.

Chunking strategy taken literally. Hmm.

Thanks for the headsup.

it seems they claim to have fixed it :

> September 14, 2026—KB5129195 (OS Builds 26200.9457 and 26100.9457) Out-of-band

> [Remote Desktop Services (known issue)] Fixed: This update addresses an issue affecting Remote Desktop Services (RDS) after installing the September 2026 Windows security update (KB5122880). In affected environments, RDS might become unstable, causing RDP connection and sign-in failures or servers to become unresponsive during Remote Desktop configuration. Related tools, including Microsoft Management Console (MMC), RDS Licensing Diagnoser, File Explorer, and the Windows Update page, might also stop responding.

When you place any piece on the board in any position, you can drag that piece all over the board to "reveal" the painting behind it whenver you move your piece.

Is that intentional or a sneak peek?

Some government bodies and officials post "official" public communications on x/facebook which may not be mirrored or oreserved anywhere else in same format.

There has to be a reasonable way to preserve these comms for public benefit.

FYI you can draw lines and triangles with multi-touch

Anthropic:

AI models should be explainable so that we can ensure and verify alignment. Responsible AI 101.

Also Anthropic:

No not like that.

How do they still purport to champion alignment and explainablity of AI if the reasoning traces are going to be hidden.

I hope in near future _that_ layer of abstraction -- looking at a fairly standard application window with minor UI variations and reasoning about what area / labels within the UI mean what (possibly paired with app documentation) -- could probably become a light-weight fine-tuned vision model it itself that can run fully locally.

I very much like the forgiving UI -- you can either go back and correct your mistakes or move on with the typing after mistyping a keystroke.

We can adapt (steal) this idea (thanks vibe coding) and practice our own topcis of interest and learning daily!

Good point.

Maybe if this philosophy / approach has a catchy name it might be more widely adopted and we all can have better more performant web and apps.

Cool.

If anyone were to propose such a thing in my country -- there would be hue and cry about national security and that this will be misused by anti social elements

How does Japan not have same concerns / voices?

Interesting

I did not know human urinary tract was suboptimal.

Are there examples of "discontinuous" evolution where a chance mutation stumbled upon a much better way to do something that allowed a species to jump the gap?

I imagine it would be very rare because chance mutations in a single individual that confer extraordinary survival / mating advantage to offset entire population would be rare like being a superman among men.

I ascribed it to the likes of Google vacuuming all this data at scale from gmail service -- which Amazon actively wanted to avoid (gifting free data to Google to train on and potentially advertize Amazon-alternatives to via personalized/targetted ads based on purchase patterns and preferences)

Anyone can take a "snap" of the phone screen with another phone's camera so this is a losing battle anyway.

It amuses me that browsers in incognito mode refuse to allow screenshots on mobile. But same browser running incognito on a desktop can be merrily screenshotted. What is the difference they are trying to enforce based purely on device form factor/OS.

This prevents us from taking scrolling screenshots (a native feature in many smartphones today that is often useful when there is more than one screen of content).

I dislike this hijacking for that reason and wish there was a way to turn it off.

Tangent

What is the current standard for generating valid BPMN diagrams. Does Mermaid support the level of control needed for formal diagrams of this nature with possible L1/L2/L3 drill down, multiple swimlanes, etc.

This conversation is important but mostly they highlight transactional and most visible examples.

I can tell you even at one higher level of abstraction, this is happening a lot --

A business analyst supposed to ask discovery questions.

A project manager supposed to list risks and assumptions and a timeline.

A consultant expected to outline a new proposed solution and the pros and cons and trade offs.

An architect supposee to present and defend a solution design -- and the change review board supposed to ask questions to validate it.

All of them are lobbing slop grenades at each other -- just packaged in normal / human-looking formats like excel and word and powerpoint and email -- but slop nevertheless with minimal to zero value add by the human, thanks to proliferation of copilot (among others).

We are hastening the rot.

This is productivity theater.

This is where some of the 10x engineer and AI taking jobs is happening I am afraid.

So maybe SKILL.md and CLAUDE.md should have model version pinned?

Somewhat like AWS CloudFormation templates start with something like TemplateVersion: 2017-10-01

But maybe that's old school thinking still.