medhir

It’s amazing to scroll through this whole product page and leave feeling like I don’t know what it really does / who it’s for.

Why are these features compelling? I went through the whole page and still don’t know what OS runs on this laptop… the value prop for this is incredibly unclear.

taps the sign: https://medhir.com/blog/right-to-root-access

It’s definitely speculation. Not sure it’s unfounded, the project has built some level of support for all other Apple Silicon Macs: https://asahilinux.org/docs/platform/feature-support/overvie...

Perhaps it is the way I framed the article, but the core of my argument is not “MacOS on iPhone”.

Yes, I do personally desire that as a niche thing, but the broader point is the nature of a locked bootloader that prevents any third-party software from being loaded / installed without getting an approval from Apple.

Whether the desire is third party apps on iOS outside of Apple-vetted channels, or entirely new operating systems like Linux or MacOS, I’m mainly arguing I should have the right to modify the software however I’d like as the device owner.

My contention is that the definition of said product and its inherent capabilities is being gatekept by a corporation that would love you to buy both an iPhone and Mac, and treat them as separate. In fact, I do have both already! But I still want rights to modify my iPhone as the computer it is.

The MacBook Neo is a great example of just how fungible these categories are, at least as far as the SoC that runs them is concerned. I paid for my iPhone in full, there is no reasonable justification for why I can’t repurpose it / modify it as I see fit.

No exactly, Apple is playing in our faces, all while people continue to defend the “differences” of device categories and the subsequent justification of shipping iPhones and iPads with locked bootloaders.

Now that MacOS has been demonstrated to run on an A-Series chip, can we please finally get unlocked bootloaders to run MacOS on other iDevices?

https://medhir.com :)

wow — this hit me hard.

I live in Seattle, and got laid off from Microsoft as a PM in Jan of this year.

Tried in early 2024 to demonstrate how we could leverage smaller models (such as Mixtral) to improve documentation and tailor code samples for our auth libraries.

The usual “fiefdom” politics took over and the project never gained steam. I do feel like I was put in a certain “non-AI” category and my career stalled, even though I took the time to build AI-integrated prototypes and present them to leadership.

It’s hard to put on a smile and go through interviews right now. It feels like the hard-earned skills we bring to the table are being so hastily devalued, and for what exactly?

would buy this in a heartbeat if I could run macOS… it’s such a shame iPad hardware has been so held back by Apple’s lackluster software strategy for this long

Avoiding iOS 26 like the plague for exactly this reason. It feels like Apple decided they need to leverage their GPU silicon for a fancy GUI update without any concern for usability downgrades.

Location: Seattle, WA

Remote: Yes, though hybrid preferred

Willing to relocate: Yes, to SF / Bay Area only

Technologies: Go, Python, Typescript, AWS, Azure, GCP, SQL, LLM-accelerated development

Résumé/CV: https://medhir.com/cv.pdf

Email: mail [at] medhir [dot] com

Engineer turned PM. I’m a highly creative, cross-disciplinary thinker and have spent time in tech doing everything from building user interfaces, managing distributed systems, to most recently leading high-impact security initiatives as a PM at Microsoft Identity.

This work included rolling out phish-resistant auth across Microsoft's developer tool products (think Visual Studio, VS Code, Azure CLI, etc) through Entra's suite of authentication SDKs, and enabling Proof-of-Possession policies through Conditional Access to protect 100k+ employees' access to sensitive resources such as source code and internal build systems.

Actively interviewing for staff / principal product roles, mainly within platform teams for B2B enterprise SaaS products. Most open to hybrid roles with the appropriate scope. Opportunity to move into a lead / management role is a nice-to-have, but not an immediate priority.

I definitely understand the intention, it still just feels a bit intellectually lazy to me.

Maybe it’s just my own aversion to using AI for creative work coming through… I can’t get myself to even add an AI generated image to a personal blog post.

something deeply ironic about this piece using an AI generated video at the top of the page.

like, of course AI is coming for culture if even the New Yorker, a very well known American cultural magazine, is willing to leverage it to avoid paying an artist to create a short video.

True, I wasn’t thinking very deeply when I wrote this comment… local models indeed are prone to the same exploits.

Regardless, giving a remote API access to a browser seems insane. Having had a chance to reflect, I’d be very wary of providing any LLM access to take actions with my personal computer. Sandbox the hell out of these things.

Personally, the only way I’m going to give an LLM access to a browser is if I’m running inference locally.

I’m sure there’s exploits that could be embedded into a model that make running locally risky as well, but giving remote access to Anthropic, OpenAI, etc just seems foolish.

Anyone having success with local LLMs and browser use?

This is pretty much exactly what Bluesky / Blacksky offer in the form of custom feeds. Gives users the ability to choose whatever type of feed they want, algorithm curated or not.

And if you’re unhappy with the selection, you’re free to create your own.

100% in alignment with this! Direct quote from the end of the post I linked:

“In the broader conversation of right to repair regulations, we also need to be thinking about a "right to root access" for computing devices.” :)

I agree there’s a ton of benefits from fully disconnecting. I’ve had periods completely off social media which were quite blissful, though I’ve always ended up coming back to some extent due the social aspects you described.

Given the choice between apps controlled by large corporations preying on attention vs decentralized protocols, I’d much rather have the latter.

I think having examples of apps that don’t demand one’s attention so fiercely, or provide better options for tailoring “the algorithm,” are an important step towards ending the mass social media addiction society is currently experiencing.

This is a great point. Not sure if it’s possible, would be great if there was some way to reclaim the notion of installing software as a general practice, regardless of whether a computer is “mobile” or “desktop”.

Like people still download software packages from the web on Windows, MacOS, and Linux… right? Maybe hard to grasp for the kids that grew up with tablets with no notion of a file system, idk

Every day we stray farther from the premise that we should be allowed to install / modify software on the computers we own.

Will once again re-up the concept of a “right to root access”, to prevent big corps from pulling this bs over and over again: https://medhir.com/blog/right-to-root-access

While I do a lot of typing, I still tremendously value hand writing. Whether that be journaling on a (somewhat) regular basis or sitting down to flesh out a concept and do some deeper thinking, I find nothing quite matches the experience of putting pen to paper.

Perhaps ironically, back in college studying data structures and algorithms, the best way I found to really grok the concepts was to write the code out by hand. Sample size of 1, but there's something about that process of having to slow down that really benefits my brain in a way that typing / dictating can't reproduce.

I feel like it’s not technically difficult to achieve this outcome… but the incentives just aren’t there to make this interoperable dream a reality.

Like, we already had a perfectly reasonable decentralized protocol with the internet itself. But ultimately businesses with a profit motive made it such that the internet became a handful of giant silos, none of which play nice with each other.

I have never understood why student loan debt is the only debt that cannot be discharged in bankruptcy proceedings.

It’s so insane that we expect 18 year olds, with brains still very much developing, to be able to make such a life-altering choice. Even though I myself understood this dynamic and decided to go to a state school with max 5k a year in tuition, I’ve never faulted others in the same situation that made less than ideal financial choices.

It’s really tragic that the US chooses to economically hamper so many people for money that is fairly inconsequential in the grand scheme of things. Everyone deserves a chance at upward economic mobility, even if they made a bad choice to study something in college with bad job prospects at a high price tag.

> Thanks to large AI models, we’re now on the cusp of user interfaces built from simple prompts.

Idk. If this were true I feel like we’d be seeing the implosion of a lot of companies’ valuations. This statement of “we’re on the cusp” keeps being presented uncritically, as if the progress in LLMs hasn’t completely plateaued over the past year.

I don’t think we’re anywhere close to replacing real humans with brains. If anything, it’s becoming clearer by the day that LLMs are not doing anything like the abstract symbolic reasoning needed to build true quality software.

I distinctly remember multiple big companies quietly letting go of their AI ethics teams in 2023 around the same time the LLM craze started to pick up real steam.

I don’t think the skeptics disappeared, they just got drowned out with all the added noise that came with the new LLM hype cycle.

ohhhh the gifts multi-tenant app authorization keeps giving!

(laid off) Microsoft PM here that worked on the patch described as a result of the research from Wiz.

One correction I’d like to suggest to the article: the guidance given is to check either is “iss” or “tid” claim when authorizing multi-tenant apps.

The actual recommended guidance we provided is slightly more involved. There is a chance that when only validating the tenant, any service principal could be granted authorized access.

You should always validate the subject in addition to validating the tenant for the token being authorized. One method for this would be to validate the token using a combined key (for example, tid+oid) or perform checks on both the tenant and subject before authorizing access. More info can be found here:

https://learn.microsoft.com/en-us/entra/identity-platform/cl...