The only real correct solution here is to send mantissa and exponent as two separate integers. It's trivial to convert between exponents for whatever math you want, it can be as correct as you want, and is unambiguous.
In the HFT space you save some wire space if you can commit to a consistent exponent for some {slice} up front (think instrument/tick-size/asset-class/exchange/feed/server/whatever/...) such that you only need to send the mantissa and your clients can have a hard coded exponent. However, in similar spaces it's often worth the extra uint32 to send a on-the-wire exponent such that things _can_ change and you aren't hamstrung later by earlier "we only need cents now!" design choices when, e.g., you suddenly need to support bitcoin/... prices to full precision. (your users will thank you when they don't have to coordinate a breaking change when you want to adjust your fixed exponent)
“After two decades of Rule 611, it is high time that the Commission review its unintended consequences that have hindered — rather than enhanced — the long-term growth of our markets,” said SEC Chairman Paul S. Atkins. “This proposal is intended to simplify market structure and reduce costs for market participants while allowing competition, innovation, and other market forces to shape the continuing evolution of our equity markets. I look forward to reviewing public comments as we take a careful, deliberative approach to avoid repeating the same mistakes that brought us here.”
Definitely could be selection bias, but every time I have seen a copy of a resume a head hunter has forwarded a potential employer it has _always_ had the recruiting firm's letterhead plastered above my content, and my email removed.
An example (I have intimate experience with) is the finance/hft space in NYC -- if you're employed at a competitive player in this space in trading/quant/engineering you will almost certainly be given a phone interview w/o question at every other competitor when you reach out.
If you don't trust the 'contact us' forms on their website it's dead simple to search e.g. LinkedIn to find their own in-house recruiters and reach out directly.
Again, if you're a new grad? Definitely higher chance of your contact going right into the trash. But the target hires are still getting called back within a day.
They can't, otherwise a significant fraction of the people they reach out to would just skip the head hunter and contact the company directly.
Same reason these same head hunters will usually strip any direct-contact details out of your resume before sending on to companies -- they don't want those companies running around them and contacting the candidate directly.
IMO, these people are all grifters and uses-car-salesman. Their goal is to get as many people as possible to use them to change jobs so they get bonuses. They provide little-to-no value add in the actual process and will actively try to shovel you toward shitty companies and dead-end roles, despite how well they dress them up.
You are ~20-50%~ cheaper (typical is 30% IIRC) in the first year of your employment if you are a direct hire instead of going through a recruiter, from a hiring manager's perspective. If you switch jobs often this compounds to make your offer chances lower as well if you're going through a head hunter (I've been part of these discussions from hiring side).
I think this is largely off the mark for most engineering teams built around roughly-aligned peers.
Sure, if your team is extremely lopsided or unfocused, and e.x. has one person from every discipline, you don't want to cross train everyone into everything else. This is a sign to reorg. Youre not asking your department heads to cross-train to other departments, your PMs/devs to cross train/...
But when you have a 1-to-2-pizza engineering team of e.g. C++ engineers and a tech lead, the lead should absolutely be encouraging this "everyone is a leader" mentality. Anything else means that your tech lead is irreplaceable and if they e.g. get hit by a bus or resign tomorrow you are SCREWED. You essentially are promoting learned helplessness as soon as the subject matter leaves your narrow areas of expertise and the "leaders" are not available to offload decisions to. The best thing a tech lead can do is encourage his workers to make him redundant -- no regular process/decision/... should ideally be blocked by their absence.
As an IC, your manager dreams of you approaching him not like "I have a problem, solve it for me", but instead "I have a problem, here is my recommended fix, how does that sound?". This would be AMAZING. You can then sync on goals/reasoning/approach/... and catch out fundamental misunderstandings on both ends. If one truly is at a fork and someone NEEDS to make a decision (really only if there is conflict as to the preferred approach) then the buck stops at the designated leader. However in most situations your engineers should be empowered to make decisions when they are confident, with review/reflection helping improve/align these decision making skills with their leads/peers over time.
Defaulting to "youre the lead, I cannot-or-willnot walk down the path unless you proceed me" is shit. Sure, when starting off it's great to have an example to follow, but eventually you gotta learn to walk the path yourself (or get off my team).
I want to be promoted one day, and the only way that's going to happen is if I can ensure I have a team of reports who prove they can survive without me (and one of whom hopefully is able to step up and fill the hole).
Because I've seen AI do it myself? It will very readily take a lofty aim from the brainstorming phase, maybe mentioned off-hand or out of context in a planning doc somewhere, and regardless of if it's implemented it will claim it's part of the delivered product.
You're looking a massive selection bias. Most people in tech are _not_ saying those things (e.g. most software engineers in my circle would agree learning to code at a non trivial level is decidedly NON trivial). The vocal elite at the top of the tech pyramids (who have a vested interest in sweeping externalities under the rug) are the ones spewing that shit.
Then your issue isn't with AI, it's with your bureaucracy. Just because your company is holding it wrong doesn't mean the entire technology is morally fraught.
If your company's goal is to generate "more useless text" they would have done it with or without AI. AI just let's the peons responsible for producing that text do so significantly faster, with some percentage loss in "quality" baked in. Are you mad their jobs are easier? Was their text once not useless and now it is?
Again, it's like saying the conveyor belt is evil because it lets us generate more useless toys/candy/guns/... and research into improving the conveyor belt should instead be going toward more valuable things. However it ALSO has those effects on EVERYTHING. It lets you produce more drugs, books, food, clothes, necessities, and yes, some useless items too.
Same with AI. Sure you can use it to spew cat pictures, but you can also use it to generate significant quantities of non-trivial useful (not necessarily bullet proof, but undoubtedly _useful_) output in a fraction of the time and/or HUMAN capital (butts-in-seats, time-on-task, ...) than before. Now, as always, value is in the eye of the beholder (which is why your C suite gets giddy at all the useless text output).
In the HFT space you save some wire space if you can commit to a consistent exponent for some {slice} up front (think instrument/tick-size/asset-class/exchange/feed/server/whatever/...) such that you only need to send the mantissa and your clients can have a hard coded exponent. However, in similar spaces it's often worth the extra uint32 to send a on-the-wire exponent such that things _can_ change and you aren't hamstrung later by earlier "we only need cents now!" design choices when, e.g., you suddenly need to support bitcoin/... prices to full precision. (your users will thank you when they don't have to coordinate a breaking change when you want to adjust your fixed exponent)