Invest your own money, let compound effect be your leverage, and avoid debt like the plague.
- Manlobbi
Stocks A to Z / Stocks B / Berkshire Hathaway (BRK.A)
No. of Recommendations: 0
https://finance.yahoo.com/news/software-stock-bull...I always found it amusing when projections for future earnings would always go up. You don't see the financial press calling this out often; in this one, the piece calls out how we're seeing earnings estimates rise for software companies, but we see some real competition from AI as an actual headwind.
No. of Recommendations: 19
It is an interesting situation. There are many possible futures.
Consider this bit...
"...earnings estimates for software companies have trended higher despite seemingly daily evidence that AI will crush future profits."
What evidence would that be?
I observe that one of the areas which LLMs are most useful in a real world real business way is increasing the productivity of programmers. Software firms employ a lot of programmers. Their costs will likely go down. So the single most certain outcome is good for those companies.
Then we get on to the many uncertain consequences.
The perceived risks seem to arise from forecasts that no company will buy or lease software any more, they'll just vibe it into existence. Outside the fluffiest end of the software ecosystem--emoji designers?--I'm more than a little bit dubious about their certainty. Automating drafting didn't put architects and engineers out of business. Heck, it didn't even put draughtsmen out of business.
Jim
No. of Recommendations: 2
Some coders have already been shed at various locales, government and private, and there isn't a strong indication that was a blip. These aren't just the sort of "Meta laid off 6000 people" sort of headlines, but rather people I know - people who have been building their own hardware for years, people who code solutions to peccadilloes, on the commuter train, on their phone. These folks were, a few years ago, getting offers of 250k+ a few years ago for even showing casual interest in a new gig.
Part of this is due to a phenomenon when the Googles and such of the world were so awash in cash that they hired people just to *prevent* their competitors from hiring them. Dry up the labor pool, on purpose. Those halcyon days seem to have gone away.
The machine has hit other industries too; a decades running fluent Russian and Chinese linguist I know well says that entire line of work is trending towards "the human does QC, the machine does the work as it is quicker and ever more accurate" where a few years ago machines were turning out comical/bad approximations of what a talented linguist could yield.
No. of Recommendations: 2
Jim:
Automating drafting didn't ... even put draughtsmen out of business.
Estimated Peak Numbers: In the early 1980s, before CAD gained mass-market dominance, there were an estimated 300,000 to 400,000 drafters in the U.S. across all disciplines (mechanical, electrical, architectural, and civil). The number now is about 110,000.
So it looks like it put about 200,000 to 300,000 draughtsmen, 67% to 75% of existing draughtsmen, out of business after all.
R:)
No. of Recommendations: 16
Automating drafting didn't ... even put draughtsmen out of business.
...
Estimated Peak Numbers: In the early 1980s, before CAD gained mass-market dominance, there were an estimated 300,000 to 400,000 drafters in the U.S. across all disciplines (mechanical, electrical, architectural, and civil). The number now is about 110,000. So it looks like it put about 200,000 to 300,000 draughtsmen, 67% to 75% of existing draughtsmen, out of business after all.
Fair enough, though it depends on who you count and what their skill level is.
If you count the number of people who are now draughtsmen half the time (about a million seats are paying ongoing Autocad licenses, for example) as 1/2 person each, the number has presumably risen. Often it's more like job diffusion than job elimination. Maybe even in areas that seem more clear cut: Microsoft Word and friends eliminated a lot of "typists", but the amount of time spent on "typist" activities has risen. Instead of (say) 5% of workers doing it all the time, maybe 70% of workers do it 10% of the time.
Jim
No. of Recommendations: 4
If you count the number of people who are now draughtsmen half the time (about a million seats are paying ongoing Autocad licenses, for example)...
Reminded me of something my business partner (Chinese/American) said when we were setting up our business - "The whole of China has one AutoCAD license".
We laughed...and paid up (2007, still works in a Win7 VM, lol).
No. of Recommendations: 0
While there is a snarky, "I'll show those big software companies" case for one copy/many installations of software (and that's harder all the time) one of the reasons South Korea got hacked so brutally about a dozen years ago was there were an awful lot of unregistered, older, copied, bogus installations of Windows that were easy to exploit.
This is a very "pick up nickels in front of the steamroller" thing. But the bad guys in cybersecurity are very, very good, and there are an awful lot of nation-states that will run rampant on those who are actively choosing to be the low hanging fruit. Which is what happens when you run old, dodgy software.
No. of Recommendations: 10
Reminded me of something my business partner (Chinese/American) said when we were setting up our business - "The whole of China has one AutoCAD license".
Ah, rings a bell. A company I co-founded was selling licensed software into China in 1981-1982. Our observation of their view of copyright was "Copy? Right!"
We were wise enough to have a hardware dongle to lock the license down, so our stuff didn't get released into the wild. But unfortunately our competitors were not so canny, and one of them sold one license...so neither of us ever sold any more.
The punchline being that we were selling CAD software.
Jim
