In general, most firms form relationships with marketing lead generation companies. i.e. you pay for customers interested in buying something, and pay a reward if a sale is made.
Don't bother spamming with FAANG, as the conversion rates are still hypothetical for many. Go to trade shows, and note how sales people operate with the public... hint, the big deals are never done on the floor area.
The sales conversion rates and tax postures will determine if this type of business is viable in your area. =3
Most people start side-businesses for tax reasons, and those that seek others to solve their problems for them... usually don't last long in business.
I've witnessed many firms run the gauntlet with varying levels of success, and would suggest the following:
1. sell what the customer already wants, as people with loss aversion stick with what they already know.
2. sell what makes customers feel good buying, and reward them with actual functional utility in their life
3. Never compete, focus on service with a novel niche product. Stupid people by their nature destroy everything around them regardless of long term benefit.
4. Never hire people unless absolutely necessary, and contract with tax responsibility clauses when possible.
5. Never buy equipment unless absolutely necessary, or lease when possible
6. Never enter legal or subscription agreements even with your own legal specialists feedback
7. Never become a poser burning $170k/month on labor in a vestigial office
8. Position your firm to leverage tax and grant programs
9. Stay quiet (especially online in a sea of cons), and only talk about the distant past when people try to goad you into telling them how you make revenue
10. Avoid bums in suits as many are dangerous well practiced thieves. Never let technical staff talk with the customers, or vendors. Some people go crazy when they see a bit of money, and do not behave rationally.
11. There must only be 1 president, and all agreements must be in contract form.
12. Never risk more than 15% of annual revenue on ANY deal. Customers lie and disappear on rare occasion... Large firms can grab your firm like a dog with a rag doll, and may still stiff you on the contract knowing the legal and fiscal power asymmetry
13. Chasing customers means your business model still needs work. If people are happy with what you are providing, than growth should naturally happen every year
14. Go to trade shows to see what other people are selling, and ask yourself what else does the customer need
15. Cash is king, as long as the money flows most other problems are irrelevant
Ubuntu Desktop 24 LTS: Kernel 6.0.8 will work on older GPU/Laptop hardware, but OS will be deprecated in 2029
Ubuntu Desktop 26 LTS will be out in a few months: Will be supported till 2038, but note old GPU drivers may not work on more modern Linux Kernels above >6.0.15
The normal Ubuntu Desktop requires a few days to make it usable, and a lot of customization to make it enjoyable. However, network printer and webcam access is usually trivial to install. google equipment installs before you buy... ymmv
Dual boot from two SSD if you need to work on the machine. You will swear less when (not if) you break something, and not everything windows works in Wine or kvm. =3
And unstable for novices that have no clue what they are working with OS wise.
The virt-manager is easy for most Desktop folks looking to drop Win11 in a frozen backing-image sandbox with a local samba folder loop-back mount (allows fake network share in Win11 or MacOS guest OS.) =3
Virt manager is the least intuitive (discounting actively antiuser crap) program I've ever dealt with. I still don't quite get it and I've used Linux exclusively for more than 5 years.
GUI might not be as powerful, but in my experience, it's similarly non-intuitive as alternatives, such as VirtualBox / UTM (macOS) / VMware Fusion/Player.
For anything more complex (e.g. GPU passthrough) you will need to drop into manually modifying XML files.
The GUI. Random permission errors, python tracebacks, saving the settings don't always work, and the mysterious charade with "storage pools" that causes new permission problems.
I just started using Incus now. It seems way more intuitive. Its remote feature is amazing too.
An unconstrained json/bson parser without recursive structure limits must be bounded somehow. In many cases, the ordering of marshaled data cannot be guaranteed across platforms.
The best method is walk the symbolic tree with a cost function, and score the fitness of the data compared to expected structures. For example, mismatched or duplicate GUID/Account/permission/key fields reroute the message to the dead-letter queue for analysis, missing required fields trigger error messaging, and missing optional fields lower the qualitative score of the message content.
Parsers can be extremely unpredictable, and loosely typed formats are dangerous at times. =3
LLM are not AI, but are a great context search tool when they work.
When people first contact ML, they fool themselves into believing it is intelligent... rather than a massive plagiarism and copyright IP theft machine.
Fun is important, but people thinking zero workmanship generated content is sustainable are still in the self-delusion stage marketers promote.
I am not going to cite how many fads I've seen cycle in popularity, but many have seen the current active cons before. A firm that takes a dollar to make a dime in revenue is by definition unsustainable. =3
I like coding AIs because they're plagiarism machines. If I ask you to do some basic data manipulation operations, I want you to do it in the most obvious, standard way possible, not come up with some fancy creative solution unless it's needed for some reason.
If I'm dockerizing an app, I want the most simple, basic, standard thing - not somebody's hand-rolled "optimized" version that I can't understand.
Fact is, the tech sector is filled with folks that find zero joy in what they do, chose a career for financial reasons, and end up being miserable to everyone including themselves.
The ex-service people would call these folks entitled Shitbirds, as no matter the situation some will complain about everything. Note, everyone still does well in most large corporate settings, but some are exhausting to be around on a project. =3
The reason we don’t have the right to be lazy is because of the people who find “meaning” in toil. I do not want to work and AI is the most anti work technology in human history.
Bertrand Russel literally wrote a book called “in defense of idleness” because he knew that heavy hitters like him had to defend work abolitionism. The “work is good” crowd is why we can’t have nice things. You guys are time thief’s and ontologically evil. May all work supporters reincarnate as either durian fruits or cockroaches.
You seem very passionate about your opinions, but are you happy?
The fact remains LLM can't reach comparable human error rates without consuming 75% of the energy output of our entire local galaxy.
While I find true Neuromorphic computing topics more interesting, the emergence of the LLM "AI" true believer is deeply concerning to those that understand how they are actually built. =3
I just had an AI write a toy game engine with realistic camera and lens simulation on the view from scratch in rust in one day while i was working on other stuff all for the price of a $20/month Cursor subscription
"AI" LLM don't write anything, but copied someones symbolic isomorphic work that could fit the expected definition in the reasoning model.
Like all copyright submarines, your firm now runs the non-zero risk someone will sue for theft, or hit the product with a DMCA claim. What is the expected value of piracy versus actual business. =3
Information wants to be free. No one in any administration now or in the future will ever go back to the "let's sue grandma for 1 trillion dollars" era of the early 2000s. Piracy is good and important for national security.
In general, people expect to see someone at least tried to find details/definitions on their own. Then to ask other people to take time to clarify subjects. In this case, LMGTFY:
By the time later Game releases were in production, world assets were already being heavily recycled. It doesn't describe the popularity of expansion/add-on packs, but rather the quality of the content discussed by the parent thread. Jokes like "learn to shoot, while walking backwards..." are still meme truisms from the Games design.
"Kitbashing" is a term originally from film/TV/model-builders for visual effects models with heavily recycled parts from multiple kits (see red dwarf or star wars set documentaries for details.) Accordingly, many modern games also mix generic re-skinned asset packs rather than hire fussy artists. The joke meme about "if you see barrels, than you know the game developers were out of ideas..." highlights how process people try to build products hoping people won't care about the drop in content quality.
YC has a population of folks that bury anything that doesn't fit personal opinions. It is a poorly structured interface in some ways, as people tend to interpret context based on whatever they were doing 30 seconds beforehand.
Also, spider traps and 42TB zip of death pages work well on poorly written scrapers that ignored robots.txt =3
reply