It's a threat. The US President does not have the right to pick out individual companies based on nothing but a whim to set individual 'tax rates' (or protection money, whatever you want to call it). The Constitution doesn't even give the President the power to set Tariffs at all. This is not the rule of law, it is rule by the whim of the head of state.

Productivity is a hugely important measure; it's just that you can usually tell who is interested by exactly which productivity measure gets trotted out.

Your ability to have goods and services is fundamentally capped by your ability to produce them or produce other stuff that you can get people to provide them to you for; so a society simply cannot be more wealthy without being more productive. However, anyone who trots out 'per person per year' productivity statistics is usually a member of team "we aren't whipping the labor units hard enough"; and anyone who fails to mention either that distribution is a...very separate...question from production and generally gets answered separately; or who ignores the fact that not working is very often one of the things that is most rewarding about being productive when you do work, is generally not speaking with your interests in mind.

I'm not sure where the exact size is(I suspect that it differs, potentially substantially, based on the margins and barriers to entry in a given industry); but there's sort of a sweet spot where a company is small enough to be hungry but big enough that you don't need to deal with suppliers going out of business all the time. Unless you are doing some "Walmart seeks lowest price for container of toilet paper" thing having vendors who will give you the shirt off their backs isn't actually ideal; since then they freeze and you need to find new ones and that's a huge hassle; but there's a distinct tone shift as companies go from having actually useful people being genuinely pleased to have your business to having outsourced billing reps say that they value your business while telling you to pay up and it will be 10% more next year.

The case for quantum computing is also pretty heavily tilted by the specific clandestine applications. If you could have one that does RSA factorization of useful sizes earlier than people think you do your spooks will love it; have a fantastic time. Maybe that is worth eleventy-zillion to you.

If you don't think that is doable, or your spook budget isn't big enough; it's much less obvious why 'quantum' should be different than 'classical' in terms of being expensive enough to manufacture that you can just buy some off the shelf once it becomes commercially viable and publicly acknowledged.

It's the sort of thing where having zero capacity to even fast-follow is probably a bad sign(though more because that means that all your physics programs are basically trash and you have no precision engineering and fabrication capability than because it would necessarily be a huge problem to quietly buy 'quantum' computers or compute time through a shell company in Dubai);but the hype is definitely a mixture of the assumption that the first mover will absolutely cash in, aided at least in part by the desire of clandestine services to, hopefully, have the first mover doing some work with them before the viability becomes publicly known.

"AI" tries to have some of that flavor of hype (with the 'zOMG AGI will become our God Machine and rule over us!' school of sales pitch); but failure to deliver on that seems to have sent it pretty sharply into the "or we could fast follow" vein as well. Especially with the established viability of 'model distillation' as a technique that can be used against hosted proprietary models as well as for internal R&D. Again, something where complete inability to do it would be bad for other reasons(you'd basically have to have atrocious access to math talent, and not even super top shelf math talent, along with an inability to design non-hopeless matrix algebra compute hardware and handle nontrivial memory size and bandwidth or fast interconnects; which would not be ideal even if LLMs are wasteful toys); though those reasons seem unlikely to be problems given continued EU presence in telcoms and CPU design.

If anything, while Europe's leading-edge logic fabs aren't desperately exciting(thought they get partial credit for the fact that basically everyone either uses or wants ASML); ARM (yes, technically Japanese now, after the famously clear-sighted UK let them be sold off for a song) is basically the god of cost effective technological fast following; and the world is drowning in ARM cores. They capture a fairly modest percentage of the value; but that's a finance bro objection. Not necessarily worth not just buying x86 stuff that Intel and AMD are desperate to sell you; but if you wanted an "jEUche" indigeounous design you'd pretty quickly have something totally usable if not necessarily commercially optimal. It's not exactly Elbrus on 90nm.

It wouldn't entirely surprise me if the ass-pulling is a collective effort.

There's the obvious incentive for the oh-woe-is-me-crushing-regulatory-burdens types to want to make being forbidden to shift risks to bystanders sound as cruel and oppressive as possible; and there's the less public-facing but more likely to be systemic interest within IT in using regulatory compliance concerns(often but not necessarily honestly) to push back against things that they otherwise wouldn't have the clout to.

The system only starts running in compliant mode because either 'the business' knew that compliance would have to be part of the objective and so didn't object to it being part of the project; or because IT was able to point to compliance requirements when someone tried to skip a step. Once it goes into production; same applies to changes and maintenance. If some sales analyst sneers that you are just an obstructionist cost center and outlines their bold plan to slam all the CRM data straight into Snowflake? Well, well, that sounds like a fairly serious GDPR issue; and it's that, not IT's relative institutional clout, that is going to get that transparently stupid idea stopped, or at least moderated.

That sort of thing produces...inherently flexible...numbers: doing careful, rigorous, IT costs more than winging it(though less than it looks if you examine a long enough period to include a meaningful comparison of the incident and disaster rates of the two styles); but if you want the number to sound much bigger you can quote the 'cost' as being the entirety of whatever ended up being done, since someone said that their winging-it-with-random-accounts-on-a-service-with-an-alarming-privacy-policy-and-no-MSA plan was basically zero cost.

Depending on what story you prefer to tell you can plausibly factor a lot of 'regulatory' costs under the "just IT having their shit together" bucket; or you can look at the IT budget as being a theoretical ideal of fast, cheap, and out of control being crushed under a 40% regulatory burden. Honestly; if you were looking to do that I'm surprised that it's only 40%. It does gradually go from 'difference in perspective' to 'outright lie' as you increasingly classify objectives the company has for its own reasons (like being able to notice embezzlement before it gets too out of hand; or not having competitors reading internal email as easily as employees do) that overlap with regulatory requirements as 'regulatory'; but there's a lot of wiggle room where you can chalk something up to either being the cost of fairly basic IT professionalism by people who care about decent work or as 'regulatory'.

This Thomas Odenwald is, according to the beigest social network, an 'independent consultant' described as "Tech-savvy, growth-oriented business executive with strong sales growth record".

Is there some reason why the article is following him like he is the provider or denier of Europe's hope for technology? It seems particularly weird when much of his experience is apparently with such noted innovators as SAP and HPE. Sure, somebody's got to shovel ERP and commodity boxes with attitudes that have outgrown their value; but that's half spadework and half sales people who look well compensated enough that customers should take it as a warning; not really brain trust of the future material.

I'd honestly be a little surprised if that specific area of policy is the problem. The American arrangement of being able to get health insurance automatically under a group rate if you work for someone else; but only as a risk-assessed individual on generally less favorable terms yourself(somewhat moderated by the Obama-era attempts to curtail the more aggressive culling of undesirable policyholders) is pretty much the most direct "Yeah, you should go work for the man unless you are young, healthy, and childless" incentive you could devise and still pass of as a 'benefit' rather than a move toward indenture.

I would expect a welfare state, if functioning, to reduce some of the really desperate flavors of technically entrepreneurial activity. If you are out at the red light doing squeegee man stuff in the hope of hustling tips out of people you are almost certainly there because you need the money badly; but I'd be much more surprised if it depresses tech-sector tier stuff. If you are allegedly dissuaded from entrepreneurship because you can technically afford to do nothing in a lousy apartment somewhere rather than seeing that as having a lot of runway(at least until you start doing something with higher costs than software development); I suspect that you aren't feeling very entrepreneurial.

They also have to show some odds of permanence, or at least long duration.

Moving a production line isn't going to be quick or cheap; so you'll only want to do it if the cost disparity looks like it is going to remain for some time. Executive orders by a mercurial child who already changes his mind unpredictably and could stroke out any time are...not exactly...that sort of assurance.

I suspect that the real innovation was on the finance side. With 'Open'AI's peculiar ownership structure and the ongoing slapfight about whether it is obliged to continue pretending to be a nonprofit or whether it is able to become a for-profit company that just loses tons of money; it was probably worth something to someone to slip a nominal 6.5 billion in equity out the door to slap the butterfly keyboard guy's name on whatever generic ODM mics-and-cameras box they end up slapping out.

No, it is a useful observation because it gives us something to look into. Just because you don't know how to negate a proposition off the top of your head doesn't mean it can't be done.

It seems quite plausible that if a LLM generates a response of a certain type, it's because it has seen that response in its training data. Otherwise you're positing a hypothetical emergent behavior, which of course is possible, but if anything that's a much harder proposition to negate if it's negatable at all with any certainty.

I think semicolons cut in the other direction. They do technically allow you to ramble at greater length before hitting a period; but they impose more structural complexity you need to think about.

From my mercifully brief encounters with the unhelpful parts of reddit the true ADHD-addled text structure is essentially a faithful transcription of a moron telling a story about something interesting only to them: A relentless torrent of short sentences without the mercy of paragraphs. Just a solid square of text with a relatively smooth right edge because most of the words are short and the gaps from things getting moved to the next line are small.

I love semicolons; how else would I legitimize what would otherwise be run-on sentences?

The national debt keeps skyrocketing and the payments on the debt have become onerous. Not a problem?

It's inconvenient to uncover the facts because so many released statistics fail to distinguish between HEV's which ultimately get all their energy from gasoline vs PHV's which could get anywhere from 0 to almost all of their energy from electricity, depending on owner needs and habits.

They also fielded a very successful formula 1 team in the 80's. For sure, the Honda of that era is gone. Like Sony with the walkman and boom boxes. Or Japan as a whole, really.