joshspankit

Also a small bug: when rotating in response to the prompt to play in landscape, the button does not show and a refresh is needed while still in landscape

Indeed. It’s very inconvenient to ‘cd` and have to do the whole permission dance to read a file

Can you make a version for VR, mostly for those “have you seen this yet” moments? Just the same data, same rendering, free movement by grabbing space (though earth pov would be a nice bonus). I’d pay $1-10 and even $1-$10/yr

> once they were famous enough & had some money to work with (to e.g. pick a more open chip).

This is where I tapped out. They could have done so much to open the industry for makers once they had traction but instead they went the other way. Disappointing at best.

Maybe I’m being charitable, but my impression was that originally they only had access to non-open gpu chips and therefore were forced to have black-box gpu blobs

If the ram chips pass proper testing, why should anyone care where they come from?

If RMA process includes hardware verification then rpi doesn’t get bled out by absorbing the cost of “counterfeits”

If they widely tell people what the market is doing and make all the tools public, they will offload the work of dealing with counterfeits while levelling up the tinkerers

The people who RMA will if you make it part of the submission process

> I think Apple's real problem is that they built machines that can last.

When they can justify it, they will add features to macOS which bog down the older systems. Eventually they will say that older systems just don’t have what it takes to run macOS and stop supporting them

Apple has never had an issue with taking machines that can last out of the market

See iOS for a great modern example

I suspect that instead of discovering that AI can become human-level by taking major leaps, we are discovering that human consciousness is actually simpler than we give it credit for

There was a time I would have agreed with this statement, but now that I’ve “seen how the sausage is made”, I believe it’s a fantasy.

Look at rowhammer: a completely novel exploit that was off the collective radar

And then, look at the software industry as a whole: an industry that works towards refined and perfectly secure code is also working towards boring and restrictive, essentially the opposite of it’s trend so far

If you have a pc to run steam, you don’t even need to jailbreak

For simplicity of explaining, I’m going to group web software parts in to “hard” and “easy” with AI assistance

hard:

- Overall security

- User permissions in multi-user deployments

- Input sanitization

- RDBMS schema design

- Parsing

- Bit-rot protection (here because it’s rarely on the radar of the general user)

easy:

- DB queries against a known schema

- Working in a front end HTML/CSS/JS stack

- Making UI components

- Moving those components around

To your question about why moving front end away from devs would reduce load: a person making a UI they want is a different category of work than creating a UI that everyone wants. Even if it’s easy, it’s distracting to be thinking about everyone’s use-cases and refining the details.

AI compute is a unique facility: can be placed anywhere, takes in mainly energy, does not produce the same volume or composition of waste, has noise pollution, and needs minimal access.

I am not a regulator but that feels like something that needs new rules

I would go even further to say if we can get zero taxpayer dollars in funding, fair and equitable land costs, infrastructure paid for by the companies, and solid ecological plans that are leveraged on threat of removal (not fines), they could be a net good where they are built

There was an app that hit it’s peak during covid: Clubhouse.

Many powerful and interesting people were having powerful and presumably private conversations in rooms because they say people walk in and out and legally CH could not record without consent. But what they could do was “collect metadata”. Emotional and tonal analysis is metadata. So is time, date, location, username (each user had their own mic stream), email, etc. Transcription is also metadata.

Essentially the tech has left the laws in the dust as you could now reasonably and legally record enough metadata to accurately resynthesize a recording.

Like prompt repetition, I wonder if reminder checkpoints stating ~”question assumptions, stay open minded” would completely remove this problem

Is there anything similar/newer that’s still a good deal?

Aha, sorry, I hadn’t realized that GOS had their own EOL timeline.

What I meant to ask was “If you use GOS, does it matter if Google ends support” and after your comment I read deeper to find the answer “yes, because GOS relies on Google to maintain the firmware”. Thanks for the push

I would be even more shocked if instead of adapting, they’ve continued to lose $10b a year

observing is the result

You missed the core of what I was saying: not vibe code a whole thing end to end (I agree, not sustainable), embrace the idea of having devs build the storage, parsers, schema, and security and let everyone vibe code the rest.

Like, let’s use a simple photos example: a photo file has metadata standards already embedded, but doesn’t store it’s edit history, event, or who’s in it. Immich already does a lot that’s great here with parsers and ML and if they changed focus to the data storage layer they could essentially boil down to a table for edit history, a table for events, a table for people, etc, the ability to store in SQLite or sidecars, and an API layer for people to hand to their AI.

Not only would it reduce the dev load, but then we could have an ever-expanding sphere of personal data ownership as Futo adds schemas/tables for video, audio, documents, and maybe everything else most people need to build their own UIs for specific uses or even “god view” UIs.

I like this article more than I like Immich. It’s well structured, helpful, and kind (not that Immich isn’t those things).

I feel like Futo is going in the wrong direction. In the age of AI, anyone can write a DB query, anyone can design a simple UI, and with docker the structure of hosting is no longer a barrier.

I think Futo and Immich should be focusing on data itself. Export, import, store, parse metadata, process, suggest relationships

If you’re using Graphene as the OS, does it matter if support ends?

What about $1/year per private repo?

Also my experience

We’re not even close to the limits of AI optimization so finding the theoretical bounds is going to have to wait

We could very easily push for giving parents the kinds of control and auditing that hackers, manufacturers, and governments currently have

I know this is hn and not a product development meeting, but:

If you’re targeting data for people to use on the same call then requires more intense work, while if you’re targeting data for people to use after or on an ongoing basis (eg an established business meeting with staff/vendors/etc) then having a predefined one seems good. Same interface as a CRM: picture, some audio clips for users to choose from, and confidence scores on each section of the recording for cases where people sound similar or are talking over each other. Over time the diarization gets better as users accept/reject/tag samples and that work helps them feel more aligned with the tool.

Since the participants are known and limited, have you tried building around samples tagged with user/person names?

The alignment of the lower stomach muscles (and them being out of alignment due to weight-related issues) is not talked about enough when it comes to low back pain