HNHacker News
TopNewBestAskShowJobs

dgoodell

250 karma · joined July 20, 2015

[username] at g mail dot com
submissionscomments
dgoodell··on Men That Don't Fit In (1911)
Are you saying that it’s impossible for the lack of persistence or focus to be problematic in someone’s life?
dgoodell··on Meta Used A.I. To Ban Accounts on Facebook and Instagram
This happened to me a couple months ago. My Facebook account was disabled, as well as my instagram account, since it was connected to the Facebook account. No explanation, and no possibility for appeal.

I had the account since it was only available to college students in like 2008. No idea what it could have been, I barely even used it lately.

I tried to make a new account a couple times with different email addresses, but it didn’t let me.

I guess I’m just banned from Meta for life, most likely due to a false positive from AI.

dgoodell··on Ask HN: Has anyone had success with SBIR grants and what is the process like?
I work at NASA and I review SBIR proposals from time to time. Most proposals are pretty mediocre. Some are physically impossible.

Coherent writing and formatting helps. Clearly explaining what your tech is and what advantages it could offer the funding agency for the listed objectives in their particular topic. You'd be surprised how many companies don't actually make any clear objective case with facts and figures, or look at all from the organization's point of view. Also, keep in mind that the reviewers may not all be deep experts in every possible obscure tech. You don't want to force them to do a ton of research just to understand wtf you're talking about.

Even if you do everything right, you might not get selected for obvious reason. At least at NASA, there always seem to be managers with their own opinions that have some kind of mysterious influence in the background. Plenty of good proposals don't get awarded, so it doesn't necessarily mean there was something wrong with it.

dgoodell··on Richard Dawkins and The Claude Delusion: The great skeptic gets taken in
Are you suggesting 85 year olds typically have more wisdom and are less easily fooled by things?
dgoodell··on Why IPv6 is so complicated
IPv4 + all the other stuff you need to actually make it work in the real world actually seems more complicated than IPv6 to me.

Maybe they’re comparing the minimal implementation on a home network. But even then I’m not sure the claim holds up.

People learned IPv4 when they were younger in a more incremental manner and take it for granted now.

dgoodell··on No Good Engineer Is Arrogant
Project managers are right until proven wrong, and engineers are wrong until proven right.
dgoodell··on Guinea worm on track to be 2nd eradicated human disease; only 10 cases in 2025
Rickover: Did you do your best? Carter: No, sir. I didn't always do my best. Rickover: Why not?
dgoodell··on AI sycophancy panic
What makes you think an LLM is internally consistent?
dgoodell··on Go Gray, Not Cray: Why You Should Grayscale Your Phone
I guess if one color pixel was significantly less efficient, and that color was also overrepresented on the display, then MAYBE changing to grayscale would require slightly less power to display the same intensity. But I don’t think that convoluted scenario probably isn’t what this person was thinking.
dgoodell··on The architecture of “not bad”: Decoding the Chinese source code of the void
Every single example given under “In Chinese, affirmation is often compiled through negation:” sounds completely natural to me, as a midwestern American English speaker.

However the direct affirmations are also acceptable. Maybe the difference is more that both are pretty acceptable in English, but that is less true for Chinese. Or at least the version he speaks.

dgoodell··on SpaceX in Talks for Share Sale That Would Boost Valuation to $800B
Option 1 isn’t really an option, unfortunately. There are no viable single launch options using it. So it’s really SLS x 2. But building and launching one SLS at a time is almost too much as it is. If that’s the only option, I think Artemis is dead and we should start over.
dgoodell··on The Department of War just shot the accountants and opted for speed
I think we use the same PPBE process at NASA. Many of the systems and procedures that NASA uses are are defense-derived. If it's anything like what we do, then it's a total mess and we mostly just go through the motions with it, knowing it doesn't actually reflect reality and it's kind of a waste of time for everybody.

However, it's risky to assume that scrapping a crappy system will result in things being better. The current shitty system was almost certainly the result of scrapping and replacing something else that had some problems.

Anyway, hopefully this works well, because we'll probably end up copying it at NASA.

dgoodell··on SBIR Mills are draining America's innovation fund
I'm a NASA civil servant and I'm regularly involved with SBIR contracts. While there are definite wins that make the SBIR program feel totally worth it, those are probably exception.

A big factor, however, is that much the tech NASA needs to develop for space applications has no commercial viability outside of NASA. We're the only potential buyer. And since it wasn't directly motivated by a specific NASA program and funded by a NASA project dollars, it has little chance to continue development once the SBIR contract ends. For many technologies, SBIR funding is the only fundamental tech development funding they get at all.

How can a private company develop a technology only the government needs when the funding from the government is extremely irregular, inconsistent, and insufficient?

I guess they could have a bunch of irons in the fire so they have a decent shot of getting some funding for anything they can do to stay in business. Maybe some of them eventually turn into SBIR mills.

dgoodell··on Staff disquiet as Alan Turing Institute faces identity crisis
As a NASA employee, I think this is extremely true. I would bet most of the second group doesn’t or even comprehend that they’re doing it.
dgoodell··on A general Fortran code for solutions of problems in space mechanics [pdf]
It’s amazing how many times I search for some space-related technical thing and I end up with a paper from NASA Glenn (used to be Lewis)
dgoodell··on 4k NASA employees opt to leave agency through deferred resignation program
NASA was originally formed as NACA to research aeronautics technology and help advance our uncompetitive nascent aerospace industry.

Building spacecraft and space transportation systems like the shuttle came later and is a very different type of task.

Unfortunately, the more flamboyant manned space flight and science missions have gradually come to dominate NASA, and much of the fundamental tech research that made it possible in the first place has been deprioritized and defunded.

dgoodell··on NASA has a list of 10 rules for software development
The mars probe unit problem was mismatched between different software programs. One was telemetry reported by the spacecraft to the ground, the other was software on the ground that used this telemetry to perform calculations that reported the results with the incorrect units.

Maybe the answer is that strong typing should somehow continue outside of the individual programs and be embedded in file formats as well?

dgoodell··on Fire the Contractors: Paradoxically adding government employees reduces costs
I work at NASA in tech development and I have found that we have a lot of poor quality support service contracts. The faceless contracting companies supply staff that technically fill roles like providing IT support services, or supporting purchasing. More often than not they are so utterly unhelpful and unknowledgeable that it's actually easier and faster to do things myself, totally defeating the purpose. The contracting mechanism seems to add significant extra communication overhead to everything making things much more sluggish, bureaucratic, disconnected, and just plain unpleasant. I actually care very much if my work gets done.

Getting a large contract in place can be a miserable slog and take a huge amount of time and effort to sort out, particularly with the cumbersome government contracting rules and laws. Good contract documents are also really challenging to write. Often times the results will be non-optimal, terms will be interpreted in ways you didn't intend, or you with you had put some more info in there. In many cases I believe the timeline to get a good contract in place can be comparable to the work that we want to perform. That's just silly.

Use of contracts for tech development creates large disconnects and significantly reduces our control and responsiveness to changing needs and ideas. If NASA employees are doing the work we can easily pivot when circumstances change and re-prioritize labor and much more quickly drop bad ideas as we learn new things. We can start investigating something without completely knowing what we're doing and figure it out as we go along. That sort of thing is harder to do with a contract. If the work is being done by a contractor, changing anything is vastly more difficult and complicated, and often not even worth the effort.

If we have a device or something developed by a contractor they often manage to contaminate it with some kind of proprietary info making it much more difficult to use and communicate the data. The tools and devices we develop internally don't have that problem and we're free to use, adapt, and communicate technical info about them as much as we want. Also, if we develop something ourselves, we inherently more deeply understand it and can more quickly make modifications or test out new ideas. That's less often the case when work is done on contract. IMO many of our most valuable developments are done internally due to the enhanced flexibility.

Not that all contracts are bad. There are plenty of cases where using contracts makes great sense and works out terrifically. However, you often just have to hope the right sort of company has decided to exist because doing it ourselves is often not an option. I have absolutely been told about a contractor: "I know they're not the best, but they're the only one interested in doing this work. If don't fund them for too long they may lose interest and then we'll have nothing".

There are plenty of other problems unrelated to contractors as well. But over-reliance on contracting is a big one.

dgoodell··on Berlin-Paris high-speed rail route symbolizes Europe's thirst for train travel
Well, there still isn't. It's not really a new line, just an indirect connection via existing high speed lines that takes over 8 hours. I don't think that's gonna be very competitive with air travel which takes less than 2 hours.

Apparently the highest average speed rail line in the world averages 318 km/h and the straight line distance between Paris and Berlin (through Belgium and Luxembourg) is about 875 km, the time could be about 2.75 hours, which wouldn't be too bad.

In reality the lines wouldn't be able to go straight and you might want to avoid going through any countries other than France and Germany. The best case scenario would probably be a little over 3 hours, which still wouldn't be too bad.

dgoodell··on U.S. math scores drop on major international test
They stated a hypothesis and asked if anyone had collected data that could support it. You responded by asking that person if they have any data. Perhaps you scored poorly on reading comprehension.
dgoodell··on It's Time to Stop Taking Sam Altman at His Word
And that’s not really the same thing. Did someone invent the LED bulb in 1920 and that cartel crushed it? Not really.

In reality, the biggest problem was they had no incentive to invest in new lighting technology research, although they had the money to do so. It takes a lot of effort to develop a new technology, and significantly more to make it practical and affordable.

I think the story of the development of the blue LED which led to modern LED lighting is more illustrative of the real obstacles of technological development.

Companies/managers don't want to invest in R&D bc it’s too uncertain and they typically are more interested in the short term.

And it’s hard for someone without deep technical knowledge to identify a realistic worthwhile technical idea from a bad one. So they focus on what they can understand and what they can quantify ().

And even technical people can fail to properly evaluate ideas that are even slightly outside their area of expertise (or even sometimes the ones that are within it )

dgoodell··on Sorry, GenAI is NOT going to 10x computer programming
So far it has definitely saved me a lot of time googling api docs and stackoverflow for me.

Anything that requires applying things in novel ways that doesn’t have lots of examples out there already seems to be completely beyond it.

Also, it often comes up with very unoptimal and inefficient solutions even though it seems to be completely aware of better solutions when prodded.

So basically it’s a fully competent programmer lol.

dgoodell··on What Is a Particle? (2020)
Are you saying that the universe cannot be represented by mathematics?

I imagine you could use that argument to shoot down pretty much any explanation.

dgoodell··on NASA investigation finds Boeing hindering Americans' return to moon
Probably the same reason the Space Shuttle wasn't cancelled even through pretty early on it was clear that it didn't make sense.

Upper management decisions are seldom made for good technical reasons.

dgoodell··on Joe Biden stands down as Democratic candidate
I used to see it more like that. But the reality is that one of those candidates is going to get elected whether you vote or not. So not voting is basically just letting other people decide for you. Which is rational if you genuinely believe that will make no difference.
dgoodell··on Will We Run Out of Data? Limits of LLM Scaling Based on Human-Generated Data
As someone who works for the nasa, I’m not so sure. You’d be surprised how much stuff gets randomly thrown away to save space.

And it’s going to get worse I now that paper files are disappearing.

I wanted some info and data from a test we did 9 years ago. It was a pretty big deal, lots of people involved, many millions of dollars, multiple nasa centers contributing. Every single person on the test randomly kept their own files for the portion of the test they were responsible for. And the only copy of the raw test data was deleted by one of them to save some space when upgrading. There is no record anywhere of what equipment was used for the test.

One of my coworkers has 4 TB external HDD that he keeps everything he has ever worked on. It’s not backed up anywhere else. It just failed and he thought he lost everything, luckily I was able to recover most of it. Wtf.

dgoodell··on AI Search: The Bitter-Er Lesson
That’s a good point. Imagine if an LLM could only read, speak, and hear at the same speed as a human. How long would training a model take?

We can make them read digital media really quickly, but we can’t really accelerate its interactions with the physical world.

dgoodell··on FAA investigating how counterfeit titanium got into Boeing and Airbus jets
I think the FAA generally performs audits, not inspections.

They usually make sure the paperwork is in order. Less likely that they make sure the paperwork is actually correct, and vastly less likely that they make sure that the actual things happening in the shop are correctly done.

I worked in an FAA repair station that repaired commercial jet engine parts. We always got the same FAA inspector every year. We never seemed nervous when he would show up.

The only auditor that seemed to really be digging to find stuff was the GE financial auditor to make sure they were getting their repair royalties.

Oh and one time an auditor for an airline snuck in and stole one of his airline’s parts, or something like that. He was making the point that we had zero access control and literally anyone could just walk into the building.

They should really start testing the employees. It doesn’t matter what the paperwork says if the employees are incompetent.

dgoodell··on FAA investigating how counterfeit titanium got into Boeing and Airbus jets
Technically it was in the fan disc, and not a turbine blade. And while there was a defect, a big part of the problem was that an inspector most likely missed a crack during overhaul. The crack was present at a previous inspection, which they knew because there some fluorescent penetrant remaining in the defect.

I worked in a turbine engine component repair shop for 8 years. We had an NDT inspector fall asleep a lot in his booth and miss cracks. I’m pretty sure they ended up firing him. But maybe not as quick as they should have.

dgoodell··on Mars rover mission will use pioneering nuclear power source
The radiators don’t have to increase in size really. If it’s more efficient you don’t have to reject as much heat.

But definitely a large increase in complexity as Stirling engines are mechanical devices and thermoelectrics are solid-state. The Stirling engines can be over 25% efficient rather than like 6% for thermoelectrics.

NASA has operated a Stirling engine in a lab for 17 years so it’s feasible, at least.

Page 1 of 3Next →