Hacker News Reader

Having helped implement NIST 800-171 for a small enterprise company preparing to work prime government contracts I can assure you it costs money either way.

Time and money spent working with cybersecurity professionals writing SOG/SOPs which map the business practices to the standard. Money spent buying and setting up some of the necessary equipment to handle CUI properly. Time (indirectly money) spent redoing decades old business processes to map to the new paradigm. Time (indirectly money) as the business spent time training their staff up on the components relevant to their business units.

I agree that building a culture is at the root of it, however, that culture alone isn't going to move the needle unless everything else is in place as well. Worse, you have to walk the company through the 5 phases of grief because in general people don't entirely understand what is actually going on and why it's important. "So you can get government prime contracts" is an easy answer, but doesn't erase the confusion behind all of the hand-wavy procedural work required to execute them.

I'm not convinced that implementing cybersecurity standards actually improves security.

No disagreements there. What's a better answer though? No one wants to do it, most non-tech people don't understand it, yet the standards and procedures need to improve to reduce the attack surface area - especially for businesses which support critical infrastructure.

You could legally mandate companies to insure themselves against cybersecurity risks.

It's really an interesting idea:

https://www.cato.org/sites/cato.org/files/serials/files/regu...

This creates a "market for effective practices", because companies that follow good security practices would get cheaper insurance.

But it would allow insurances and companies to figure out dynamically what is cost effective and what isn't.

There would be lots of financial incentive to figure out what protects you effectively and what is just security theatre.

Most companys already do. Unfortunately for the insurance companys their actuarial tables are backwards looking and cyberattacking is one of the fastest growing industrys.

The smart insurance companys are all getting out of cybersecurity insurance because the premiums on the policys they wrote 10 years ago when the average attack cost $1,000 to remediate do not look so good when the average attack now costs $100,000 to $1,000,000 to remediate. Since the expected probability of a successful attack in any given year that requires payout is approaching 100% these days the insurance needs to be priced something like 10x-100x the prevailing prices to be survivable.

Eventually it will shake out once the cyberattacking industry becomes mature because at that point backwards looking projections make sense. Until that time, we are in for a wild ride.

This is intensely defeatist. Regulations and compliance are essentially what you described and they work for banks (despite the clever responses I'm sure are incoming). For the critical industry sectors they need to be buying the 50k appliance. And they need to be looking at a checklist that they only want to complete to avoid fines. Culture will accrete later.

I disagree with the GP’s approach of defunding education to improve education, but I’ve seen the culture that accretes from checklists and expensive security appliances.

Everyone involved correctly comes to the conclusion that the company’s security policy is purely performative.

The people that care leave. The people that don’t care about securing the system implement the useless checkboxes and put the standalone box in the correct room (and might even plug it in), then get promoted.

FIPS security certifications are the poster child for this. There’s basically no way to get a securely architected system certified, so you end up adding vulnerabilities and (in the best case) a configuration parameter like the old “enable FIPS certification (default off; do not enable unless you are required to do so)” button from windows.

50k appliance doesn’t do shit if people work from their personal laptop and use shared passwords.

The only thing a regulation that mandates a 50k appliance will do is boost the appliance makers sales.

An appliance also externalizes responsibility. I don’t need to think of anything, because we have the box that goes, “ping”.

It emphatically does not work for banks. They spend hundreds of millions of dollars a year and any bank CISO would tell you, off the record, that their systems could be completely breached for less then a million dollars. That is not because they suck at doing it, that is because commercial cybersecurity is just plain garbage. That is literally the best you can get no matter how much you spend and no matter how “good” your people are if you use commercial IT security practices. Nobody does better than that who uses anything from a big brand cybersecurity vendor.

The only reason banks are not totally wrecked by cyberattacks is that:

1) They actually are being constantly attacked by ransomware and the like, they just do not disclose it; I do not remember the budget line item the payments all go under but it is one of those compliance/administrative ones such as for insurance.

2) Cyberattacks, contrary to their standard portrayal as always being run by “state-sponsored” groups to make the companys being destroyed look better, are largely run by kids. They are largely unsophisticated on financial and business matters so do not know how much to ask for, how to maximally leverage a breach as a first party (i.e. do not know how to use say investment bank deal details to do effective trades), and how to actually cash out large amounts of money undetectably.

3) The non cybersecurity processes and fraud detection are actually fairly effective. There are large number of accounting checks in place that make it very hard to do simple hacks (like make the numbers in one account higher) that can not be detected. Banks have the equivalent of very complex, bespoke DRM systems that requires knowledge of financial systems and of their internal processes to crack. As I said before, the hackers are largely financially unsophisticated, so it is very hard for them to do this.

No, it is intensely defeatist to just give up on actually solving the problem and demanding useless security theater in the vain hopes that it will somehow improve the situation.

Regulations and rules force banks to be held accountable. Whether or not their cybersecurity prevents all breaches is besides the point. They are incentive in minimizing breaches, and when one occurs they're the one who have to cover the losses (yes I'm sure you're already typing clever examples of banks avoiding responsibility that one time but the fact that we rarely hear about then, as you say yourself, is that they cover it up and refund all the losses themselves).

> demanding useless security theater in the vain hopes that it will somehow improve the situation.

The security theatre required by these new rules might not prevent all breaches, or any breaches, but they will force utilities to be accountable for their failures.

No longer will it be "Oopsie, mean hacker hacked us. Not our fault. Whatchugonna do anyway? Your business had to shutdown because no power/internet? Tough luck I guess".

Now it will be "Oh fuck, hackers got in. We'll have to cover our customers losses and spend millions in PR to spin the story and avoid total bankruptcy."

Do you really not see how that's still an improvement?

That’s not what regulators do in this space. Instead, the rules are of the form “use certified vendor, and you will be shielded from liability”.

Also, the vendors and the people being shielded largely write the rules.

No, liability requirements force banks to be held accountable. They do eat most of the costs of the direct consequences of being hacked. Luckily for them, the people hacking them are still, frankly, inexperienced from a business perspective so their losses have to this point been largely immaterial.

The losses are not low because the cybersecurity processes and regulations are good, they are low because even though the vault is being guarded by the metaphorical equivalent a chihuahua, all they are doing is stealing the pens because they do not know how to pawn the gold in bulk.

This state of affairs is changing very quickly. 15 years ago the hackers were walking into the vault and asking for $1,000 because they were 16 year olds who thought that was a lot of money. Now the good ones are 30 years old and are still walking into the vault, but asking for $10,000,000. Last I saw the number of attacks was increasing 3x year over year and the average ask was increasing 3x year over year.

The entire cybersecurity insurance industry is already underwater at current premiums. I have heard they have basically stopped issuing policys over a few million dollars because it is literally ruinous. It is no longer becoming possible to paper over the omnipresent security deficiencys with insurance, betting on the incompetence of the attackers, and betting only one of you is going to get eaten.

The sensible solution right now would be liability requirements on advertising security. You can not insinuate you are “secure” whatever that even means. You can instead only advertise a dollar amount per customer which is tied to a mandatory bug bounty. If you are a big bank with 100 million customers, you can put up a 10 billion dollar bug bounty on breaching your systems and then you can advertise $100 of security on your customer deposits.

This would force them to put their money where their mouth is while not preventing new small companys from existing as long as they truthfully report that they are not providing security. It would also help us pull the customer desired security requirements forward in time instead of waiting for the destruction first.

Absent that, what is going to happen is that everybody keeps lying about their security to trick their customers into trusting them. Then one of them is going to get hit by a truly expensive attack and all the customers will be caught with their pants down.

The only technical solution to this is either dropping reliance on these easily hacked systems, or throwing out all the existing crap that was never designed for security and can thus never be retrofitted to be secure such as Microsoft, Linux, Cisco, and Crowdstrike. Instead we must deploy systems designed for security such as those targeted to conform to the (now deprecated) Orange Book Class A1 or Common Criteria EAL 6/7 systems which are certified and proven to resist nation state attackers such as the NSA.

> They are largely unsophisticated on financial and business matters so do not know how much to ask for, how to maximally leverage a breach as a first party (i.e. do not know how to use say investment bank deal details to do effective trades), and how to actually cash out large amounts of money undetectably.

I wonder how many actually sophisticated attackers gain access to banks and other companies and purposefully keep a low profile but skim off low-bandwidth information meant for internal communications only, then aggregate and sell the intel to hedge funds or trade stocks on their own.

Maybe the reason why we think that most attacks are caused by unsophisticated actors is because those are the one who are actually detected.

Just kind of an interesting hypothesis that I’ve heard variations of before.

Regulations and compliance is not what is being proposed here.

Regulations are rules, this is a formalized incentive structure. They're a tad different.

Regulation is good to a degree; there's a balance to strike between overly burdensome regulation and regulation which basically may as well not exist. The lines between these are sometimes vague as the times evolve and are also subject to regulatory capture by interest groups (political groups and companies).

The incentive structure generally makes spending money free or more free so long as an entity abides by the guidelines set out by the incentive.

You need both and both need to reflect the correct goals for the times.

> This is intensely defeatist.

It's realistic.

> Regulations and compliance

Which might be nice. "Do this, or you will be penalized like this." These are marketed as "incentives" which is "Buy something, anything, and here's some abstract additional cash."

> And they need to be looking at a checklist that they only want to complete to avoid fines

The problem here is, you do the checklist, and you get an incentive. I wonder how that's actually going to be implemented?

All I see here is a government subsidy for security vendors. I'd rather see more liability and insurance requirements for security breaches, including personal liability for executives. Whatever it takes to get past checkbox-driven security.

And schools are just a subsidy for the textbook industry.

Regulation & incentives are literally the tools governments have to shape policy. It's a start, but moving the needle on issues can be hard, slow work -- especially when it seems like so many people have the attitude that if a particular action wasn't the their perfect, preferred solution, then it shouldn't have been done at all.

In your analogy, the vendors are the people manufacturing textbooks that are mandatory, but never get opened and cannot be resold on the used market. The regulators are the administrators (or whoever) that require the useless textbooks.

Of course, there are many useful textbooks with none of the above properties, but they don’t require complicated incentive structures to ensure they’re made mandatory.

I worked at an electric utility in the tech side and we had a steady brain drain of guys going to Google, Meta in order to 2x or 3x their TC. Money talks.

https://www.google.com/search?q=US+electric+industry+profits

That seems to be a "management is unwilling to pay enough" problem, not a "the industry is struggling and needs a government handout for relief from the burden of enhancing security" problem.

Congresscritters are demanding we cut "fat" out of basic social assistance programs by increasing requirements from beneficiaries but we're going to give electrical companies, making half a trillion in profits running basic infrastructure, a big handout, with trivial or no requirements?

Pass.

You're looking at revenue numbers, not profits. Profit in the utility industry is 0 on operations and maintenance expenses and 0-9% (depending on whether you are a public utility or co-op, of which there are many, and whether your regulator thinks you've been good or bad lately) on capital expenses.

Expenses must be approved by the regulator as well. How strict that is depends on the state but in New York for example it's approved at the department level. All utility salaries are below non-utility equivalents and keeping people from going to Google or Meta by paying higher salaries is effectively illegal.

This is not to boo-hoo for utilities. The ones that are for-profit do fine. But they are not parasitically extracting vast profits because that isn't allowed.

Money isn't everything, but it's really hard to run an effective school or cyber security enterprise without money.

This is lost to many in the West. There are things you cannot buy or strong arm your way out

Well that's a strawman. I've seen churches, corporations, and families being run on amazing shoestring budgets, but you couldn't tell from the outside, because the holder of the purse strings was judicious about spending.

Years ago, I learned a peculiar thing about corporate or municipal budgets: sometimes a line item has to be spent or it won't come back next year. So there is often some strong impetus to spend $XXX,XXX -- or else -- and the deadline comes on June 30, and often that is how you arrive at bad spending decisions, because it's "burning a hole in your pocket".

That's not what the word strawman means.

"it's really hard to run an effective school or cyber security enterprise without money."

GP is actually understating this; if my budget and revenue is $0, then I cannot run a school or enterprise at all and it will file for bankruptcy and close.

A strawman is the informal fallacy of refuting an argument different from the one being discussed. I never said that a budget should be $0 or "without money". That's a reductio ad absurdum.

The government only has so many levers it can pull, though. Even if we, the enlightened of HN have some idea of which investments are beneficial (for example, hiring a good CISO instead of buying the $50K cybersecurity box), laws and policies move too slowly and are too hard to change once enacted.

So this situation still seems better than a rule which only benefits companies that hire Cisco Certified Security Posture Experts or something. There's at least a chance that some of these companies have competent CTOs (or equivalents, at smaller utilities) and will spend in the right area.

Government has plenty of levers to pull - they just don’t pull them.

Want to have utilities (or any industry) put cyber security front of mind? Just tell them that they are liable for consequential damages for any service outages related to a cyber security incident, and watch how good they get at cybersecurity.

Government absolutely does pull those levers. I think people underestimate how closely utilities are regulated. It's a monopoly after all. The government tells them annually what their profit for the coming year will be.

The state public utility / service commission determines how much profit a utility is allowed to make (really it's return on equity) based on federal guidelines. Every year the utility must make a rate case that includes what it's going to spend and why and why its shareholders deserve to get anything on top of that. If the utility has misbehaved that RoE number can literally be zero, though this would be rare.

Punitive profit slashing (not quite that extreme) has definitely happened for major customer-facing IT issues and would probably happen for a major breach.

> You can't just purchase a "deluxe cybersecurity firewall appliance" and plug it into your network, even if it costs $50,000

Those cost a quarter of a million per year nowadays...

Don’t you need money(investment) to hire personnel, training, etc?

Never heard of a hardware vendor accepting payment in the form of company culture.

So just be smart about how you spend your money? Like any other department?

It’s not about smarts it’s about incentives, if I’m incentivized to buy cybersecurity equipment but there is no incentive or penalty for not actually preventing cybersecurity breaches, a whole bunch of stuff is going to be bought with little effect.

It's a utility company - someones uncle booger works there - there will always be zero culture there face palm