Why? A dozen different reasons. Of course LLMs are one facet, there's also the commodification of software and infrastructure which means buying something off the shelf is far cheaper than running an engineering org in-house, there's also the fact that the market is extremely oversaturated with software engineers with hundreds of thousands laid off over the last few years, then there's the aggregate effect of advancements in PL and software system design which makes it a lot easier to do more with less, the broader homogenization of runtime systems with modern browsers and mature cross-platform toolkits... and many many other factors. All these trends are converging on downward pressure for demand, and I personally don't see any reason why the trend will reverse.
A coincidental combination of economical conditions happened before the layoffs, and I know correlation doesn't imply causation, but these conditions look like a big cause:
* Companies hired like crazy in Covid * Section 174 got disabled * Interest rates rose
This made money much more expensive, and employees became a much higher cost due to the fact you hired like crazy, so you have a ton more, and you can't amortize them, also combined with fears of recession in 2023.
In a very short term, this cocktail of conditions made operating a company much more expensive, thus the layoffs and reorgs as an attempt to cut costs.
What you are saying is also true, but I see that taking effect over a longer period of time.
Company A hires like crazy due to Covid, then everyone else does the same because they must be on to something!
I agree, I don't know why they expected covid conditions to be the new normal.
I think the only company who didn't get on the crazy hiring bandwagon was Apple.
> The domestic research or experimental expenditures . . . otherwise taken into account as a deduction or charged to capital account under this chapter shall be reduced by the amount of the credit allowed under section 41(a). Read in conjunction with Section 41(d)(1)(A), discussed above, it seems that all taxpayers claiming a research tax credit will necessarily have costs which are treated under Section 174A and thus subject to the reduction specified under amended Section 280C(c)(1).
> To our knowledge, many taxpayers have interpreted this language to mean that there is a reduction under 280C(c)(1) only to the extent the research credit exceeds the amortization allowed under Section 174, generally 10% in the year the expense is incurred under the applicable half-year convention. In that case, there would typically be little or no reduction to deductions and capitalized amounts, and correspondingly no reason to elect a reduced credit in lieu of a nonexistent or minimal reduction.
https://www.morganlewis.com/pubs/2025/07/new-section-174a-re...
TL;DR: I don't think we're out of the woods yet
So in "a few years" (let's say 2-3 years) you'd be able to make between 5 and 10 years' worth of European net pay. If you don't raise your spending, that will easily cover your living expenses during the next recession, even if you spend all of it unemployed.
And that's if you start now. If you've been doing this for 10 years already ... wow.
Besides that, if we assume AI will take many jobs in the coming years, what are you to do if you suddenly become unemployable with your current skill set and need to, for example, go back to school, or find an entry level job in another field with a potentially significantly lower salary? Live off of your savings?
Is there someplace I can find information about how section 174 aligns with the frequency and size of layoffs?