HNHacker News
TopNewBestAskShowJobs

johngalt

5,758 karma · joined July 21, 2009

submissionscomments
johngalt··on Understanding Cybersecurity Frameworks: NIST, ISO, and More
NIST CSF = The encyclopedia which breaks security down into as many areas/steps/sections as possible. If you are planning a 500-person security department, this is how you give them all something to do. The idea is to accomplish the task with manpower rather than elegance. CSF itself is mostly just a pointer to NIST 800-53. For truly large-scale operations it can be an ok fit, but for most organizations it is overkill that your cyber-insurance vendor will still expect you to do. Otherwise, best used as a reference not a guide.

ISO 27001 = Not super familiar with this one.

COBIT = The management/process focused version of NIST CSF. Great if you have an executive suite CTO, CISO, CRiskO, CPrivacyO, and want to coordinate their efforts in a program that divides responsibilities among them and associated committees. Includes maturity modeling, which gives it a +1, but it is distant from anything technology related. Instead, it is all about which committees should be formed to decide on risk management strategies etc...

PCI-DSS = You'll do this one because VISA makes you do it. Much more actionable than NIST or COBIT, but it depends on the third-party auditor who is issuing your attestation of compliance. "Your label maker has it's default password?" = audit finding.

CIS18 Controls = The most actionable/lightweight framework now that they have incorporated maturity levels (aka implementation groups). Not as thorough as NIST or COBIT. Well implemented, CIS18 is enough for most organizations provided they do not have a specific security standard or requirement in their industry.

johngalt··on Why there are so many cybersecurity vendors and where do we go from here
The proliferation of security vendors is similar to the proliferation of weight loss clinics and gyms. There are plenty of fads, with new businesses popping up to either chase or create the interest. The people buying these services desperately want something which can plug into their existing habits without significant changes.

Similarly, the solutions for cybersecurity are simple but not easy. It involves operational and administrative discipline. Businesses which lack this discipline collide with security problems and spend a great deal of money downstream of this problem. Vendors sell what businesses want to buy, not necessarily what is most effective.

johngalt··on Ask HN: I have 176 logins/accounts. How many do you have?
It's a solid case for FIDO2 keys.

- Inherent MFA that is faster and more secure than SMS codes, or app notifications.

- One pin code and/or biometric to remember rather than hundreds.

The answer isn't to add another layer of management to passwords, but to eliminate passwords as a method of authentication.

johngalt··on The socialist calculation debate
The problem with planned economies is mismatched incentives, not computational capabilities.
johngalt··on Poste.io – Complete Mail Server
Is there a reason that forwarding is the path chosen for this rather than setting gmail as your MX?
johngalt··on More than five whys and “layer eight” problems
An excellent write up of the changes in perspective which occur with experience. No organization is optimal and neither are we. The key is to incorporate these details into planning. For example if the organization can't set durable goals, don't start a 3 year plan.

Don't passively accept it, or become fatalistic about dysfunctions. If there are achievable changes to make, try to make them within your sphere of influence. Do what you can to produce positive changes, but don't allow frustration to lead to cynicism. Think of layer eight problems as engineering constraints like any other.

johngalt··on U.S. moves to bar noncompete agreements in labor contracts
Difficult to defend "as practiced". For example trying to enforce a non-compete for sandwiches is ridiculous.

There are obvious scenarios in competitive industries where losing key people to competitors creates risks for everyone involved (not just the losing employer). The problem is when employers expect the individual employee to absorb the costs to mitigate that risk. Or where the non-competes are applied overbroadly. Rather than ban non-competes, it would be better to require compensation for the term of the non-compete. This prevents frivolous applications of non-competes, which is the actual problem.

johngalt··on Ask HN: What will everyone do if everything is automated?
Until we have AGI, I expect automation to be similar to the past. Where productivity increases along with material wealth and standard of living. After AGI, "what jobs will people do?" is probably the most uninteresting part.
johngalt··on Rackspace Cloud Office suffers security breach
To anyone impacted by this. Here are a few things you might want to check, other than the obvious user mailbox migration.

0. Set your MX record TTLs as low as possible. Even if you aren't ready to change over yet, it will lower propagation delay when you are ready to cut over.

1. Lower your DMARC record TTL, and consider changing it to report only (p=none). If you keep it strict (P=reject), be sure to check your reports for rejections after any changes.

2. Aliases in rackspace aren't in the mailbox list export. This isn't just Bill instead of William, but also info@domain.com or helpdesk@domain.com etc...

3. Allow lists/block lists can be exported from rackspace, and imported in Microsoft but consider starting over from zero if you can. Allow lists are risky.

4. If you are using any advanced email threat protection tools such as Barracuda or Mimecast, they can queue email for re-delivery once the new MX is up.

My email is in my profile if anyone has specific questions.

johngalt··on Cyber Guidance for Small Businesses
I doubt CISA believes that technical and cybersecurity experience is irrelevant. This is their way to say "put someone in charge of it."

Two reasons for this:

1. The failure mode for most SMB operational risks is "no one was behind the wheel. No one thought it was their responsibility." If someone is clearly identified as responsible, they can set the basic guidelines that most people already know should be done.

2. Once the term "responsible for X" is on the table, it will tend to push the business towards hiring skilled personnel. Precisely because no one wants to take that on. Recommend that a business hire a skilled IT security headcount at market rates, and all the stakeholders will vote no. Ask a business "who is responsible for IT security? Who will handle an incident or breach?" And they will hire an IT security person after no one steps forward.

johngalt··on The flying wedge
This can certainly be cronyism where an exec brings along an entourage of sycophants to build their own empire on the company's dime.

It can also be a team selected from their broader network of known/proven professionals. People who they already know can work together and produce results.

To steelman this tactic a bit. Put yourself in a new exec's position. You were brought in with specific goals to achieve, and a tight timeframe to show performance. At that level there are no second chances, and no excuses. You've been working long enough to know a number of people who can execute. It will be months before you know what the current team can/can't do, and hiring unknown people is always a dice roll. Early team changes are expected. Conversely if you wait months sifting resumes and building up the existing team, the opportunity to make significant changes will gradually evaporate.

What would you do? Wait and see? Or put a team together of only the top 5% people from all the places you've worked previously?

johngalt··on Ask HN: Skeptical about my company going “full serverless”. What am I missing?
How to derail this dynamic early:

"Does this solve a problem we have?"

"Of the top five things that we are trying to build, does this add any of them?"

Often the problem isn't that the new methods/tech is bad but that all the effort spent transitioning to it could be better spent directly attacking the goal in the first place.

johngalt··on Ask HN: “Contact Us” Pricing
There isn't a hidden price that is only shown in the sales funnel. But rather the price changes depending on the customer.

'Contact us' = 'give us enough information to accurately price the product for you.'

Best way to shortcut the process would be to work with a channel/var that has a majority of the items in their portfolio.

johngalt··on A global house-price slump is coming?
A slump is likely. Or specifically a correction followed by a period of slow/no growth, but I don't expect a crash similar to 2008. The last housing crash was primarily due to a combination of overleveraging and overbuilding. The cost of nails, wood, and labor hadn't increased to the same degree.

The recent asset price increases have been supported by broad inflation. I don't see how prices drop significantly without monetary deflation. If we see deflation at scale, then everything is in trouble.

johngalt··on Avoiding Unqualified Gatekeepers
This anti-pattern is predicated by two underlying truths.

1. The gatekeeping behavior is individually beneficial even if they are globally detrimental.

Blocking access costs nothing, and may benefit them as people pay the toll to gain access to the resource. They may not be Machiavellian schemers. It could just be mismatched incentives or a departments only way to deal with the 'externalized costs' imposed by other departments. For example the account manager may guard their customers because an ambitious sales team is incentivized to aggressively cross sale, and the sales team does not get dinged for souring the customer relationship.

2. The gatekeeping behavior may be globally beneficial even if it is annoying to those who must go through it.

Raising the bar for access to a valuable resource keeps that resource from being squandered, even if it is more frustrating to use when needed. It is why we have to break the glass to get to the fire extinguisher. Anyone who's been in a position where they are the 'valuable resource' knows that there is a constant cacophony of poorly thought out requests, meetings, proposals, demands, etc... And without some means of filtering, or pushing requirements back to the requestor, valuable time is wasted.

Designing an organization to avoid T1 gatekeeping involves identifying why this behavior benefits the individuals involved, and following that thread to find the root cause. Directly discouraging T1 gatekeeping without changing incentives is the management strategy of Sisyphus.

Most organizations could benefit from an increased and more formal approach to T2 gatekeeping. It can go too far, but I've observed many teams held back because everyone is informally asking one or two experts for all the answers.

johngalt··on “Durable Computronics”
Quality standards are certainly required, and the software/technology ecosystem is brittle in avoidable ways. However, I'm not sure we can simply roll forward with methods we used to build bridges and use them to build software.

I expect that we will see quality management going the other way. Deriving new and more capable ways of dealing with the complexity of modern technology stacks. Then those methods backpropagate to how we manage other complex systems. As we saw with the 737 MAX, in many cases the "engineer signing off" is merely a figurehead/scapegoat. Could one person realistically know the details of every system (to a detailed engineering level) in a modern airliner? Even if they could, is that the best approach?

Our methods for durability are already overextended. Technology will be the area that makes us admit it, and provides the tools to manage it better.

johngalt··on Be critical or be corrupted
It is the inherent conflict between authority and accountability. It's unfair to be accountable for something we have no control over, but if we have control over it, we can unduly influence it. There is no bulletproof answer to this problem. Fixes are mostly about allowing for the possibility of success.
johngalt··on 2000-Watt Society
Aiming at the wrong goal. A "10kw society" on solar or nuclear is better than "2kw society" on coal.
johngalt··on 2000-Watt Society
> wonder how the average US household is at 12KW

If you've eliminated transportation and heating that is #1 and #2 for most people's energy consumption. Third place would likely be air-conditioning, and you've eliminated that too.

johngalt··on “I survived the TCP transition” (2013)
To this day I think ATM was an interesting approach. Virtual circuit switching with quality of service capabilities designed into the protocol. If nothing else, it is a great example of a complex and optimized protocol losing vs a ubiquitous and simple protocol.
johngalt··on Slow Travel
Air travel is already quite efficient. If we break it down per passenger, the gas mileage is around 50mpg. Cutting total flight distance only cuts the amount of time that the aircraft is in it's most efficient travel regime (high level cruise). With no corresponding change to the amount of fuel used to climb up to cruising altitude.

Adding extra days to every trip is not cost free. Mode switching for travel is also not cost free. Creating gigantic airports at the nearest points between continents etc... Large scale, expensive and difficult to coordinate changes; which are trying to squeeze more efficiency out of an already optimized system.

Don't mean to sound overly critical, these are interesting ideas on systemic change.

johngalt··on Nuclear is back on the table for a green future
The near future looks like Solar + Natural gas. Driven entirely by economic factors. It is cheap to build PV, and Natural Gas plants cost less than sufficient batteries/storage for PV.

I used to be very bullish on Nuclear. The numbers just don't pencil out in the US under the current constraints. We would need a vast reduction in the overhead to build a Nuclear plant, or a vast increase in the costs of the alternatives.

johngalt··on Mygov is a ‘huge disappointment,’ key architect despairs
Imagine trying to turn around an organization like that. Must be an interesting challenge.

The sheer amount of technical debt, legacy systems, dysfunctional team processes and culture. Not to mention the sheer motive inertia needed to change anything in that environment. Moving in any direction will have 1000 other things breaking/popping up to steal momentum. A Gordian knot impossible to untangle.

johngalt··on Dude, where's my fuel?
Prometheus should be laser focused on being first to market in commercial quantities. There are people who want carbon neutral fuels today even if they are more expensive than fossil fuels. Capture that entire market with v1 of the process even if it's $10/gallon. Take everything learned to make v2 at $7/gallon and v3 at $5/gallon etc...

The goal of $3/gallon is pushing Prometheus down the rabbit hole. Waiting for the perfect factory, with the manufacturing methods, to produce the perfect machine that will immediately go into large scale production, and operate on an automated basis. I expect the company aiming at 10-7-5-3 will reach 3 there faster than the company aiming at 3 to start out with.

johngalt··on Using LEOs and GEOs
Why was latency such a killer on throughput? Was it TCP window sizing?
johngalt··on Why Vendor Software Has Holes
Software vendors have two modes.

1. A partnership/consultancy where they sell a platform that is customized for a small number of large clients.

2. Solving one area well, then trying to cram as many customers through the fixed solution as possible.

The author is more critical of the second type, but both are completely rational approaches. Obviously anyone can listen to customer feedback, but the economics of software strongly favor a volume approach rather than treating every customer as a development partner.

johngalt··on Tell HN: Make sure to configure DKIM/SPF with Fastmail
Not unique to Fastmail. Any domain that sends email should have DKIM/SPF/DMARC. SPF is quickly becoming irrelevant, but it is an easy configuration item.

Recommend mxtoolbox for validating configurations https://mxtoolbox.com/

Specifically send a test email to ping@tools.mxtoolbox.com and it will advise you of your current settings.

Dmarcian has good resources on DMARC specifically, and can act as an RUA report reader as a paid service. https://dmarcian.com/alignment/

johngalt··on Tell HN: Make sure to configure DKIM/SPF with Fastmail
DMARC: Set is as p=none and read your reports from the RUA tags.

Once you are confident that all the legitimate mail is aligned, then go straight to p=reject. Many will recommend quarantine, but it's better to have an email bounce back immediately vs silently get lost in a spam folder. Outside of troubleshooting there isn't much use for P=Quarantine in DMARC or '~all' in SPF.

johngalt··on Ask HN: would you trust a company's container images or repos after a breach?
There must be a qualification process for anything that will be put in contact with a production environment and/or sensitive data. What a container validation process looks like depends on the type of organization and it's risk appetite.
johngalt··on A 2030 Self-Driving Car Bet
> I suppose it just needs to do approximately as well as a human.

From a practical/logical sense this is true, but I'm not sure it would work out that way. People would never accept FSD cars with accident/injury rates comparable with humans. FSD would have to be at least 10x better than human drivers, maybe more. Psychologically we are much more risk averse to when not in control. It's why people are afraid of elevators but not stairs, despite a 50x difference in safety.

Historically computers go from "barely able to do X", to "better than any human could ever be" very quickly. I'm with Carmack on this bet.

← PreviousPage 2 of 34Next →