mstolpm

Thank you. Have seen this after I posted.

Seems to be a warning for Macs without a built-in keyboard only. Getting it on my Mac mini but not my Air. Seems to be a good idea for that situation. Bet there are more casual users using bluetooth for keyboard/mouse/trackpad than usb.

Terrible onboarding: Why do I need to create an external account even before seeing what the app has to offer for me? Immediately uninstalled the app.

There is already the opposite situation: "available on iOS AppStore (except in EU)". And that might pop up more often in the future.

You can already "autofill" Firefox (and other) input fields with keychain data in Sonoma. Its a bit cumbersome, but right click into the field, select "autofill" from the context menu, then either "passwords" or "contacts", search/select/confirm the data to be inserted.

But I would only know about the result for my sales in some weeks or months. I have to trust the LLM in the meantime - and your example implies that the LLM answer could be proven wrong. How is that different from asking a complete stranger about better sales strategies for my business? Or even throw a dice?

EU regulations are very messy and DMA doesn't seem to be different. There are so many loop holes in the regulation that the "intention" and the effect are in part disjunct.

See the GDPR and cookie laws: End users should have been protected, but big companies (not only Meta et al, but e.g. German media corporations) are finding ways to "comply" and screw the users at the same time. Cookie consent is a mess. On the other hand, small businesses are struggling to comply and fear getting fined or getting an "Abmahnung" from a competitor.

The new rule that requests electronic invoicing (ZUGFeRD) is a welcome step towards more digitalization, but requests a human readable PDF with an embedded XML containing the structured invoice. The regulations require a human (!) to ensure that the PDF and XML match. Anyone who has ever worked with XML knows that a freelancer or small business outside of IT services doesn't have the knowledge and ability to do so. Again, they have to rely on external tools and services that not only imply costs, but themselves don't guarantee compliance.

Now the DMA. Regulations are welcome and needed for big players, but again, the rules seem to be arbitrary and full of loop holes. This is the EU and it already shows that the DMA leads to more fragmentation, both geographically (inside EU vs outside) as well as in the device space (iOS regulated, iPadOS not?). The relevant parts for AppStores ("Marketplaces") don't even mention the end user, but other competing market actors. Apple, Google, Meta all are interpreting the regulations different and of course "comply" in a way that favors their bottom line - of course, because the EU market is just a small portion of the revenue and the share holders would be angry if management would act different. And the end user already sees a growing enshitification in the EU: No more click through from Google search mini maps to Google Maps (no disadvantage for other maps providers). Browser choice menus with browsers nobody ever heard of. Requirements to accept new TOS without understanding what changed. Requests for "combining data" across services for "better experience"... But there is no mention of, for example, free side loading in the DMA, and no clear definition how to interpret the regulation.

We had all this in the EU with "Windows N" editions: special editions of Windows for European countries with unbundled components and more "free choice". Nobody wanted them: neither customers nor OEMs. And they had more bugs that the regular versions and were hated by Microsoft as well. And we have seen the flaws of GDPR and cookie law in recent years. Oh, and we have seen huge fines for large corporations as well ... and years and years of fighting them in courts, often reducing the fines significantly.

Moreover, the DMA mostly targets American "gate keepers" and this language already shows that the regulation is protective in its nature. It's interesting that even many US HN users hope for a strong reaction of the EU against Apple in particular, despite nobody has ever clearly shown the Apple has a monopoly in the EU or that customers were harmed. But what would the reaction if the EU would truly ban Apple from business in the EU or demand a split up? What would be the retaliations if the EU really tries to have the upper hand on how US corps can operate? And what would be the reaction of EU citizens if e.g. Apple, Meta oder Google would face a ban?

Try creating a new user profile (keeping the old untouched). I had the same problems on my M1 MBA and assumed complex hardware or OS issues. But all problems are gone on the new profile (still existent on the old one when used). Strange.

On the topic of Hackintoshes: They are becoming more and more obsolet with Apples focus on Apple Silicon, Neural Engines and so on. Not spec or price wise, but regarding feature parity and Apples OS support for the Intel architecture. Lots of old „Hackintoshers“ abandoning the hobby.

Why is it that LLMs are so often compared to employees and their responsibilities? In my opinion, it is an employee that actively USES the LLM as a tool and this employee (or his/her employer) is responsible for the results.

Hasn’t Google denied repair of Pixel watches still under warranty with cracked screens lately? Cited in the Verge: „… we don’t have any repair option for the Google Pixel Watch.“ [1] shouldn’t they improve their own RtR options before demanding to change Apples existing? (I agree part pairing isn’t great, but Google fighting for RtR is double speak.)

[1] https://www.theverge.com/23874281/google-pixel-watch-cracks-...

Thanks. I'm still a bit confused about how many tokens for example your example video uses up. You draw the body and it generates a first iteration, then you draw the eyes and it seems there are two new generations, then some more for the background. As mentioned elsewhere, the token system (and potential subscription model) is a bit of a showstopper for me, but the concept is great.

Another input: Perhaps I have missed it, but you should mention export options and sizes of generated results on your website. Another question that pops up is the usage rights of the generated images.

Great project, congrats.

Can you detail how tokens are used? Is that per generated version, per final (?) image, per time, export, size?

Another feedback: I initially had the wrong impression that the left side was the "input" and the right side the "output" and wondered it the app gave initial renderings the user could "uglify". The video of course cleared that confusion. But perhaps more people are like me "reading" the screenshots from left to right as in input -> output? From a UX standpoint, drawing on the side side of the screen for input of course makes a lot of sense. Perhaps the website could label an example with "our input" and "generated result"?

Much depends on the topics you cover and the target readers you want to reach. As others have noted, ePub is the standard for ebook publishing. But if you want to publish a printed edition as well in the future, your toolchain would have to support pdf as well.

Pandoc can convert almost everything in everything and is great for terminal nerds. However, be aware that formatting pdfs for common self publishing distributors like KDP (Amazon) has some strict requirements.

For someone without the nerd dns, Vellum is a great software on the Mac that formats ePub and pdf as required for different vendors and print formats. The best feature is that you have one manuscript and export to ePub and pdf. Including graphics works quite well, but sometimes clever positioning for pdf is a bit tricky. Be aware that Vellum is a paid software and it is not suited for equations and docs that require very formal formatting (on the other hand, ePub ebooks are generally not a very good choice here).

But before you choose your toolchain, be aware that writing and formatting the manuscript is only a small part of self publishing. You need distribution and marketing, proof-reading and a business model.

Amazon can help with distribution, but you could publish via your own website, Gumroad and numerous other platforms. Do you want to only sell the ePub? Or sell add-ons, courses, bundles etc.? Amazon is okay for mass market topics, however there is a large competition and you have to have a marketing strategy. Selling via Gumroad allows for bundles and greater flexibility, but is better suited for niche topics. Again: You have to do the marketing.

Does your target group prefer printed books? Amazon/KDP can help here with print on demand. Beware anything where you print (and buy) hundreds or thousands of books in advance: You will have your garage filled with them and never sell them all because of logistics.

So, please don't only focus on the toolchain and the doc to ebook step. That is one of the easier parts in the equation.

Apple wants their users to use iCloud storage. Transferring photos and video by cable is not the suggested way of usage for most iPhone users. Using iCloud sync, your photos and videos are already waiting for you on your Mac. At least, if everything works and you aren't a heavy "Pro" user. Of course, that is not only a "pro-consumer" tactic, but a "pro-service-revenue" one. And if you are a "Pro" user with lots of data to transfer, you can either but the USB 3.0 cable or subscribe to the new 6 TB and 12 TB iCloud storage options. Or just do both. Of course, you need a Pro model ... but aren't you a Pro user in that case?

In EU, WhatsApp, Telegram and Signal are way more popular. Lots of friends do own an iPhone, but seldom or never use iMessage. The first they do is installing WhatsApp. This is in part because WhatsApp had a head start in Europe and was cross platform all the time. I would not surprised if Apples claim not to have 45 Mio _active_ _monthly_ users will hold.

Vienna is full of international organizations from UNO to OPEC: https://wien-io.diplo.de/iow-en/international-organizations

Doesn't Apple have this step already? I'm talking about the online store with optional pickup in your Apple Store. I've ordered online whenever I exactly know what I want. No sales talk, no waiting. Pickup in store was fast and hassle free, but that might depend on store and time. Best of all: I know exactly that the item I want is available for me.

That would not necessarily ban AI generated content, only bad storytelling.

A physical key to your front door is a physical object and something you own - but you can still make copies of it. Someone else can even make a copy from a photo of the key. Google authenticator was (until recently) the only app I know that made it very hard to copy the OAuth secret to another device, lots of other OAuth apps made it a feature to sync/backup your keys.

I can't think of a lot of security factors (besides perhaps biometrics) where copying is really impossible. Its just not economic.

I'm wondering: How can I be certain that the model contains any personal data about me (or someone else not famous)?

For a public figure, of course there is lots of information in the training data, all public data. But when asked about me or my brother, ChatGPT either refuses to answer OR hallucinates the hell of it. Then, nearly everything is wrong and the output resembles the answer to a prompt like: "Create a short bio for a fictional character named xx, living in yy and working as zz." (Okay, often yy and zz are wrong either.)

Requesting to delete these hallucinated facts seems quite stubborn and ineffective?

No. It is anchored in the right to be forgotten and the right to claim your stored PII data regarding the DSGVO to protect companies from being flooded by requests from a malicious actor or even a competitor. Without that, a company could theoretically forced out of business by the work/cost the requests generate because not answering a request is by itself a violation.

Why are you discussing OpenAI with ChatGPT? I’m honestly interested.

I would imagine that any answer of ChatGPT on that topic is either (a) „hallucinated“ and not based on any verifiable fact or (b) scripted in by OpenAI.

The same question pops up for me whenever someone asks ChatGPT about the internals and workings of ChatGPT. Am I missing something?

You can blame a worker it something goes wrong and you trusted their result. But you can't go to YOUR boss and blame ChatGPT if YOU used the wrong prompt and trusted the result of the AI.

Salary isn't for delivering an untrustworthy result, its for being responsible for the result. AI doesn't take responsibility.

Stable Diffusion is trained on 2.3 billion images. It would take a human ~220 years to decide if an image should be included in the training set under the condition that a decision takes 1 second and the human works for 8 hours straight year-round.

GPT is even worse in that regard, it was literally trained on the whole web plus other sources. I'd be surprised if a judge would follow the argument that the training data was make by creative choice instead of raw crawling power. I'd argue the size of training data makes it impossible to call for a creative selection process for the data.

Isn‘t it kind of ironic that we complain about ChatGPT not being „creative“ enough on a thread complaining it fakes academic papers? How can the same bot be limited to only produce „facts“ and at the same time would be useful to generate great „fiction“?

But I completely agree with you: it’s often spitting out boring stuff. And yes, it makes almost everything up based on language statistics and similarities.

Perhaps, we are just in a hype cycle? There are lots of specialty tools and bots available based on GPT-3 customized for different tasks. Complaining that ChatGPT or Bing+ aren’t equally suited for every task gets tiresome. Especially if more than 2/3 of the examples simply follow the pattern „junk in - junk out“.

What would be the information the oven can expose over thus a channel? How often I cook? How much energy I use for cooking? The oven will not even know if I'm at home or not if not used. I can't think of any critical spying done that way?

Of course if the oven had cameras and microphones - but then again, the bandwidth of a covert channel would be rather limiting.

Not upgradable. But "minimum usable" all depends on usage - for a lot of users, 8GB / 256GB is sufficient, especially in office or education.

No. macOS is a tool. It might not be the best, but I get my work done with it. Linux (first question: what flavour?) would be a tool as well. Maybe more flexible - but it would need a lot of time and commitment to customise und familiarise with. I'm not paid for this time. And it is very questionable that a change would lead to much better productivity. So: no.

It's the same as a hammer in my shop: Of course there might be a better hammer out there for the current job. But I would have to browse catalogs, pay money, familiarise with the new toy. Lots of time and effort without gained productivity. No, that 10+ year old set of hammers I own does the job fairly well, thank you very much.

I don't have a citation, but a personal anecdote: When I was in charge for a website, an elderly woman contacted us and asked, if we could change the black text to a dark grey. She mentioned that the contest was too high, resulting for her in seeing ghost images of the letters. I changed the text color from black to something like #eee or #ddd (at least a change barely noticeable) and she later contacted us to to thank us, so it must have had an effect on her. So it seems, for some people with limited eyesight, contrast can be too high as well.

However, I agree with the generell sentiment that in modern web design, contrast is often too low.

Congrats on your launch! But honestly, your site gives me some huge red flags regarding your service:

You don't have a legally valid "Impressum" on fruits.de: It points to fanbase.com and that one isn't even mentioning fruits.de. So, as a German startup, it seems you're at least bending German and EU law by not providing the information necessary for proving services on the domain fruits.de. And as a result, as a potential user, I can't trust you with my business.

Moreover, I couldn't find information about your "sales tax" claim. You know better than me that selling digital products in the EU and across borders can be kind of tricky, depending on what you are selling (e.g. ebook vs software), if you're selling B2B or B2C, and depending on the origin of your customer. As a user, I'd need much more information about how you handle this and the bookkeeping you mentioned, but it somehow looks like you're relying on PayPal as the payment gateway and are not handling sales tax / USt by yourself?

And no: As someone looking for a simple solution to sell my digital products, I don't want do book a consultation for answers to the questions above.