Is there survivor bias in being stuck? We have accepted that people who leap out of their comfort zone end up always successful. But a lot of time this is not true. I know few people who have regretted jumping and making impulsive decision in the name of growth.
Good point. Nothing is without risk, and it doesn't always lead to success. Every jump, impulsive or calculated, has its own set of outcomes and consequences, and it's important to consider both sides.
Wow this is super impressive but does somebody know a way to generate consistent characters with stable diffusion?
What I mean is if my first prompt is a girl talking to cat and second prompt is girl playing with that cat, I want the girl and cat to be the same in both pictures.
Is that possible? If so any links or tutorials will be super helpful to learn.
You can do this on Dashtoon Studio. They let you upload just one image and train a consistent character Lora. It's a software for AI comic creation. Found this video on their youtube https://www.youtube.com/watch?v=EEQwEvKQGvE
Lora is by far the most versatile because you can get your character consistently in any pose and in any camera angle. IP adapter replicate too many traits from the input image, and you can't choose what not replicate like the pose. So getting a character from a portrait input to do anything else can become. For Reactor you need a generated image into which you can swap in a face. Works very well for realistic images, for stylized images the style is not maintained. Also hairstyle won't get copied.
So Dashtoon is the most reliable thing and easy thing I've found so far because collecting 20 images of a new character is hard and the properties of the images in a Lora training set are really important like how many close ups, how many expressions etc.
Check out https://scenario.gg - they let you train your own LoRAs on custom images of a character (you need around 20 or so images from different angles for good consistency). A bit simpler, and actually still pretty decent is IP-Adapter, which they also support. Having the cat be consistent is going to be challenging without a custom LoRA I reckon. See this for guidance: https://help.scenario.com/training-a-character-lora
AI researchers can make any claim, the risk of getting busted is close to none.
Didn't work ?
Well dataset was different
Didn't work ?
Well code was different
Can I try your work ?
Well it's proprietary / I don't have access / We shutdowned the cluster
But the result is guaranteed increase in salary and job opportunities.
Since these companies are publicly listed, they are by definition encouraged and encouraging to make grandiose claims in order to make themselves more attractive to investors, and they can blame the individuals if it becomes discovered.
My favorite being that Bard (PaLM version) is sentient, but it was too big this time.
Imagine a large pharmaceutical company claiming they can cure very important diseases, but the results cannot be independently verified, nor audited.
It’s ok in the short-term, but not when you make that claim during few years.
The point is probably the implication that it'll be pushed to android as a native feature (in the photos app or similar), thus it making sense for investors reading this to put money into google rather than e.g. stability or openai etc. The people writing the article are likely shareholders etc.
Serious question but does anyone get any value out of these threads? Most of the time it just devolves into hundreds of comments with links to random projects hoping to get traffic.
I think to make it more worthwhile people posting here please write a little about your tech stack, why you made it, what are your struggles, and tips for other founders, etc.
Around these parts "ideas are cheap" is often repeated, but I've failed to come up with a marketable idea for the past ten or so years.
My hobbies & interests are too niche and the problems I have in life can't be solved by tech, so I have yet to run across an idea I'd be intrinsically-motivated enough to pursue.
With that being said, I'm hoping I'll run into someone else's idea which will help me see through the kind of blindness which prevented rsync users from seeing Dropbox as something worth building, so I find exposure to these "little" ideas useful since reading through threads like these is somewhat like speed-dating for startup ideas.
This is the kind of content I come to HN for. If it makes you feel better, this kind of thread also serves as a lightning rod that contains the self-promoting of projects so you won’t see as many posts of this type.
I'm relatively technically inclined so the "tech stack" used is not really all that interesting. I don't really care about what React widget was used to create a customizable overlay text on an animated gif meme, I care about how the person found an audience and managed to monetize it.
GitHub, Reddit, "Show HN" or other areas of the internet are much better at highlighting interesting projects. This thread is specifically about monetizing small to mid range projects, so the focus is on how to acquire a meager income stream both in targeting audience and monetization strategies.
The best responses in this thread, in my opinion, are the ones that talk about how they managed to get to $500/month by identifying what problem people would pay money for, how they found customers and the specific type of transaction (purchasing something physical, subscription, one-time removal of watermark, etc.).
Yes, I love that thread. It helps me with brainstorming on new ideas.
Also, it's pretty nice to share with the small team I'm part of. We're currently working on custom client projects and we'd like to build our product. Seeing how people do it is a nice morale boost, especially for a team that lacks experience in building.
I do. I don't care about the tech stack, but I'm always curious about what got people started building something. What opportunity did they see and what led them to that point.
Yes. These are the posts on HN that I enjoy most. Tech stack etc are also somewhat interesting, but not that relevant since for most people is best tech stack is the one they are familiar with.
Hmm, I was looking at the extended keyboard layouts previously had doubled Windows keys; that's a pretty clear spot to put it there. Looks like a lot of laptops don't have the doubled-Windows keys?
Depending on the size of the keyboard and on the size of your hands, using the left modifier for that shortcut might need two hands instead of just one; it's very convenient to be able to quickly lock the desktop with only one hand.
I just tested and,on a full size model m, I can still stretch 2.5 inches after reaching the left+win from my pinkie and l with my thumb. And I have average hand span for a man.
What do you use instead? Most other browsers, especially the big blue and yellow one, provide even less control. I am always annoyed that Firefox won't let me skip an invalid cert warning though, while Chrome always will by typing "thisisunsafe" into the error page.
How does one learn about this stuff? I learned about basic networking in college (the TCP layers) etc but people doing such stuff sounds like Greek to me.
If I want to learn more about what the author is doing, is there a resource like a udemy course or YouTube channel you guys can recommend?
Back before cellular data coverage as ubiquitous and cheap, getting on an using shoddy wifi was an essential skill for anyone with a laptop. The ability to lock onto a stable wifi router, piggyback on another network, or syphon some bandwidth from a network you shouldn't ... those with such abilities managed to book tickets or hotel rooms within seconds of a flight being cancelled. Or, if you were really evil, you kicked everyone in the terminal off the router, hogging all the bandwidth for yourself. At school I had a script to randomize my laptops MAC ever few minutes then reconnect to the school's network. Mine was the only laptop that never got throttled by the wifi police for using too much data in a given session.
I used to have a little wifi antenna on my car. Some called it "wardriving" but I called it being able to check my email while traveling.
Back in the day you needed to know this stuff to get things working. It was not that uncommon for routers not to have DHCP so you had to input the IPs by hand. I think this may have been even the default and DHCP had to be explicitly enabled (e.g. in Linux installing and configuring dhclient).
Also it wasn't that uncommon to expose a computer to internet through the router, so you had to make sure that computer didn't change its IP.
I think having to set these up yourself is the best way of learning them.
The vagueries of DHCP can be learned best with a home lab, IMO
Back in the day, setting up random hardware or VMs on an isolated subnet taught you everything you needed to know about low level network protocols like DHCP, STP, BOOTP, ARP, RARP, and how to sniff it all with wire shark when you weren’t getting a lease
Containers have largely hidden this plumbing from us at a test/dev layer
I'd suggest a book. I was pretty happy with "Computer Networking: A Top-down Approach" by Jim Kurose. I find it more appealing that it starts with the upper layers (http), because I was more familiar with them.
“You can't buy a hard copy of the 8th edition, but instead can rent (and then choose/pay to keep the hardcopy if you want a hard copy book). You can rent a copy or subscribe to Pearson+ from our publisher, or rent a hard copy or purchase a Kindle version from Amazon, or rent a hard copy from VitalSource.”
Nice to see the name Jim Kurose here. Many years ago, I learned networking and C programming in his computer networks course at UMass. Such a great teacher and a real breakthrough class for me in understanding not just networks, but low level systems programming, computer architecture, and other things tangentially related to networks, I'm not surprised to hear his book is good.
Don't know about video courses, but Internet standards like DHCP are open and available on IETF.org. There are explanations on Wikipedia too. Also, if you're into Linux the HOWTO section on tldp.org can also be a big help for more practical stuff.
Get yourself a router that supports OpenWRT, install that on it and figure out what every configuration option does. Bonus points: setup WPA Enterprise on it and a DNS resolver.
Most computer networks 101 class teach how their state machine is designed only.
If you want to know "how to use them in real world", some universities has courses with "System Administration" would be more suitable. or learning the certificate program (CCNP, CCIE, JNCIP and others) materials with their lab.
What do they teach you at university nowadays??? When I was in university studying computer engineering (almost 20y ago) we were taught networking a very thoroughly! Not only every technology involved but even the small details of how everything worked (from Ethernet cable based network using csma/cd to WiFi). We even studied in details ALOHAnet!
Back in the day you would just follow The Linux Documentation Project's HOWTOs and set up each kind of software on a little computer on your home network and play around with things. Today there's nothing like that unfortunately. HOWTOs were abandoned for individual blog posts written for one specific use case at a time, and I know of no index of such blog posts.
I learned a lot at my student job in the university IT department troubleshooting why I could boot from my Xserves NetBoot server in one room but not another.
The rest I learned in the last year by switching to pfSense/Opnsense for my router/firewall.
I would take courses oriented towards getting your CCNA (Cisco Certified Network Administrator). It covers the basics while also teaching you a bit about navigating through a Cisco switch.
