3,107 karma · joined June 21, 2011
Find a good available (!) .COM domain: https://comsensei.com
Domain busy.io for sale, $25k minimum.
Formerly a Microsoft Product Planner in Redmond, now self-employed, enjoying making things people need.
Happy to connect with anyone from HN on LinkedIn in hopes of that AI-tained feed improving with some actually interesting content:
https://www.linkedin.com/in/mittermayr
I worry about all the con/pro-sumer hardware makers, from Philips Hue to Alexas, from the SumUps to the camera makers, from Netgear to TP-Link. All their products are packed with open-source libraries. And I am 100% certain that most of their dev teams do not spend time scanning these for obscure injection vectors.
And, I just had a look, here’s an article by someone (not me) who seemed to experience the exact same threat: https://killdozer.medium.com/facebooks-shame-o-gram-8e71106c...
So, while I hate saying this, be warned, I was warned and ignored it, and eventually, no matter how tall you feel, that call may come quicker than you think. Keep a second domain around ready to go, make sure you can rebrand. Just my word of (super sad) advice.
Looks great though!
I am not sure how my $8 of Spotify money subsidized Taylor Swift and/or how that is a problem? I would have $0 otherwise, not $10. Spotify enabled me to get to the $8. Fair is a hard word when the distribution power of Spotify is not accounted for, I'd argue. Again, I know nothing, it's just one perspective I often find myself in, and it feels weird (being someone who benefits from Spotify, i.e. the $8 instead of $0).
Well, you could simply bypass Spotify and just buy the music directly from the artist, no (which you indicate you do)? So Spotify isn't necessary for this, I'd guess?
Spotify isn't a payment app for artists (again, I feel super weird defending them, I am not a fan myself!), it's a distribution channel.
Well, as a musician myself, I often feel a bit conflicted reading statements like these. After all, as uncomfortable as it may be, it is <insert big pop artist> that brings _most_ people to a platform, not the indies, typically. If I upload my music to a platform that pays me 100% of the royalties, I will get to keep the most of 0 dollars, because nobody will be there to listen to it.
I would argue that the issue is mainly relevant for indies who are past the "much-needed exposure" phase, starting to blow up, and now don't get a good (fair) share of it. That would be a great moment to go off platform, and take the audience with them, but many can't (won't) do that, because it may mean less exposure. So, after all, Spotify seems to bring something of their own to the table (a pool of listeners) that they want to be paid for. To get more exposure, they need bigger artists, to get those, they need money and crooked deals to convince the big labels to license it. So, while unfair at first glance, comparing a Taylor Swift deal to what I'm getting off of Spotify perhaps (I'm only guessing), I am happy to accept the deal for the random exposure I am now enjoying because of it.
I feel like I am betraying all artists just saying this, but I keep coming back to wondering why everyone feels like they deserve the exact same deal like a big-name artist (who bring their own audience I am benefiting from).
It's the same problem most indie devs and startups have (without VC-infused Adwords budgets) — getting their product in front of other people.
I have been consulting in a self-employed capacity for 13 years now, next to doing other things, and 100% of my leads have come through friends, exclusively. This is not to show how great my friends are (debatable), but how I have been completely unable to attract business outside of my network, even in 13 years!
Marketing yourself to even be considered a point of contact for whatever it is you're offering is - in my opinion — by far, the hardest nut to crack.
I'd love to hear suggestions by others who've been successful on this front, especially how you attract out-of-network leads. It seems next to impossible at times, except for those rare (serendipitous) moments when someone reaches out after a newspaper article or being featured online somewhere.
I also wanted to mention that clearly knowing the hypothesis to test for at the very beginning isn't all that easy, in my opinion. A lot of the time, you need to walk down the path, even walk it for a good long while, until you hit the right angle or context to truly see where the project could land. You (hopefully) start these things with a "good enough" reason, but it often needs a lot of shaping and refining to get it all back into the light, for a wider audience.
Curiosity in learning new things will (should) always be a good part of what fuels the effort, it's a safety net (for me as a developer anyway), shielding me a bit from ending up with purely wasted time once it's clear that it didn't lead anywhere. At least I learned something new along the way.
Provide various templates for both pre-instructions as well as post-processing prompts. Like, some "tested" prompts that ensure (as best as possible) that the output is in certain formats (JSON, a list, a restricted CSV set, etc.), or that the input will ensure (as best as possible at least) to prevent basic jailbreaks from the main prompt.
It would take me a long time to get up to speed to what people working with ChatGPT every day have already figured out to work best in warming up GPT with a prompt as well as ensuring that the output doesn't escalate into something unexpected. Having those (reliable) templates would be fantastic for anyone starting!
Nuances, sure, but every single time I ask someone about their Moka pot, they have a slightly different way of using it than I have.
Case in point: I fill water up to just touching/under the valve, then use a medium-coarse grind with dark espresso beans, with enough coffee to fill the basket level, no tamping brushing off or touching. Then heat it hard on gas if available, leave the lid closed while it builds up but open immediately with the first bubble and put it off the heat. Let it continue for another half minute and then pour. And you?
Even if you take AI aside, it remains a fact that nothing will ever stay as it is, things will always move, the world around you will always change (we get older, everything gets re-evaluated constantly as we age). I thought we'd reached the ultimate end-goal with Turbo Pascal, then, Borland Delphi — why use anything else, ever? This is not a particular fault of the AI era, it's just one that hits you and your profession a bit closer. The exact same thing happened to SO MANY OTHER PROFESSIONS over the years.
There was an article a long while back about a US law firm that decided to try and outsource a huge case by shipping tons of boxes to a specifically-trained "warehouse" of workers in the Philippines, knowing that their law (and capability to speak/read/understand English) is similar enough to that of the US. And, from what I recall, it turned out that while this was a fantastic business decision which saved them a lot of money by managing to outsource the most expensive part of the those huge cases, it also became clear that "being a lawyer" may have lost its long-held prosperous perspective (at the bottom).
I say this because that article was written over a decade ago (if I remember correctly), and even today, law-schools are still packed, lawyers are still making a lot of money, everything is still... sorta fine (for them).
You will NOT live forever. So the question is only if you can sustain your career long enough to make it through and then enjoy your ACTUAL LIFE once you're reasonable safe, financially. I am very, very confident that you will be able to do that, even with the rise of AI.
I look at all the mid-to-large-scale projects I've been involved with, as a developer, and 80% of the real work is figuring out how to approach certain things, and how everything should tie together. Writing the code itself was never the problem. Never. For AI to take over completely (aside from writing certain functions or boilerplate starters), I'd argue we've still got a very long way to go.