Also possibly concerning, is the financial incentive to push conversations into areas where the ads are discussed. People dont always talk about consumer goods, but with the ad model you might have the situation where AI keeps bringing them up anytime it is remotely relevant. turning conversations into commercials, and missing opportunities for learning generic words by replacing them with named brands.

You: "Hi how are you?"

AI: "Thanks to my Frosted Flakes and cup of Folgers that I bought from ShopRite this morning, I am feeling great! Have you tried Frosted Flakes?"

Reality: "I'm good, you?"

that said, there could be some value in seeing brands you recognize around the words they represent (like saying "frosted flakes cereal" above instead of just "frosted flakes") since the context will help you remember what the generic words mean.

Also dont underestimate the issues you might have getting Brands on board with this. you are asking them to give you permission to use their brand name in conversations that they cannot approve of in advance. that's very tricky too

Jane sends unknown message to John

John fits the target profile of topic C

John gets shown ad for topic C

---

coincidentally, Janes unknown message was about topic C. Likely because she too believed it to be applicable to John.

---

Going further...

Maybe topic A was trending in Janes area. Jane's friend Sally got ads for topic A. sally talked to Jane about topic A. Jane searches topic A

A trend emerged where people who searched topic A, very soon afterwards searched topic E.

Jane is very late to hear the news about topic A. She gets shown ads for topic E. Jane thinks advertisers are reading her mind because she was thinking about topic E but didnt talk to anyone about it.

Yes they are ultimately responsible for all the same border crossing paperwork and such, but that is a lot easier when everyone you work with is giving you information in convenient formats and returning your calls quickly. Not to mention the ability to leverage work from past shipments as reference.

a small player in this space has a significantly harder time since they have less reference, and vendors are more likely to treat their business as a nice-to-have instead of a want or must-have.

And thus the reason for all this regulation.

In fact, nuclear reactor to be built have a higher barrier of 'prove of safety' then almost anything else.

In fact, nobody in the US ever died because of civilian nuclear reactor research or at least not in the last 40+ years.

Additionally, the public is relatively unaware of how nuclear plants fail. If people at nuclear reactor research labs suddenly stop showing up to work - what happens? People kind of assume it will explode, or become a radiation hot-zone rendering the local area unusable for a hundred+ years.

And considering the experimental reactors being discussed need parts replacement due to corrosion, what happens if those replacements dont happen?

The dangers of human inaction seem much higher for nuclear than other things. With standard fuel sources, if people stop showing up to work then power simply stops being made - that's pretty much it. And while they are at work, the process is pretty simple and robust (compared to nuclear) with a lot of room for error.

I know I've been surprised (in a good way) to learn about some of the safety mechanisms existing within nuclear reactors, but I still only sort of understand what anything means that I read - and the safety mechanisms that gave me some sense of relief is based on a lot of assumptions I had to make about the way nuclear works as a lay person.

how about nuclear scientist stop saying "trust me bro" and more aggressively educate people on nuclear fail-safes? People have trouble voting for things they do not understand. And educating the public on something they are either uninterested in or incapable of understanding is an unfair burden to put on nuclear scientists, but they are the only ones qualified to do so.

the ball will move a lot faster once lay people can exchange stories about nuclear safety that go beyond "we barely use nuclear, and no one has died yet, it's actually really safe because ... reasons? scientists said so?"

Data right now is typically bought and sold with an expectation that most of it is crap. it's faster to buy and process 5000 dirty items that probably has a few good leads buried within it than to find leads manually / naturally or broadcast random advertising. (I left the industry in 2020 and my NDA expired in 2021)

Data quality is typically assessed at the "Does this data field have a value for this line item" level. That means data vendors are financially incentivized to make shit up about you as much as they can get away with. think about it for a second, these companies are selling themselves as the source of truth. the actual accuracy does not matter, and the better you are then the less data your customers buy. the data goes stale faster than the accuracy of the data becomes relevant

Did you like a post about a fresh baked baguette that had #french as one of the 100 tags associated with it? congrats, you're french now. it's not exactly this ridiculous, but you get my point

there are some verification focused services - like they take a list of emails and check if they are valid email addresses. Some use fine print to say they are only validating whether or not it is of valid email address FORMATTING, and make no claim about whether or not the email will bounce. verifying if the email address actually belongs to the person it claims to is not part of the deal.

it's nearly an impossible task, because you have no actual source of truth to verify it against. So data vendor A and B give you different results for the same search - now what? you have to manually research and see whos "right" or "more recent".

even if it looks like good data, it might be stale. For example, company size, revenue, C level email addresses, etc all change over time.

so if a customer wants cleaner data - you basically charge them to pump the dataset through Mechanical Turks or upwork or something to have people try to verify things manually. Datasets can be large though and this gets expensive, so it tends to be better to just buy the crap data for cheaper and figure it out yourself

I have a conspiracy theory that these verification services are behind a lot of the phone spam today. they are just checking if your phone number is valid, they dont actually care if you answer.

Exactly this. But they can get away with basically anything. Worst case for them is they show you a premium ad you aren’t interested in. Best case is they guess correctly

I want to make a "game" which is more of a game environment sandbox. something kind of in-between game engines and actual games. I want to be able to create the feel of different game types like FPS, RTS, top-down adventure, side scroller, whatever - within a certain set of consistencies of my choosing. Like something to toy with different ideas for game mechanics in different settings while not needing to explicitly manage things like what a wall or door is each time i change things (things that would be defined as part of me creating the "game"). I want to be able to invite a friend into test environments as well (so multiplayer needed). I am expecting to need to make edits to the game mechanics outside of the game, possibly even through adding extensions to whatever game engine i use - so i will not be looking to set up a UI for tweaking settings. Just something that makes it easy to make changes to how the game functions, easy to spin up a dedicated server to run an instance of the created game environment, easy for another person to join that environment.

Advice needed is: is this scope feasible for 1 person and what game engine should I look into for trying it out? I am not trying to do anything particularly fancy; no overall game design, no story, minimal amount of textures and assets. Literally just an environment to toy with game mechanics.

I understand there is so much below the surface of what I am trying to do that I dont understand, i.e. I have seen an example for all the decisions that need to be made about doors, how interactions work, size, how changes are managed (this door, or all doors?) etc.

Basically, I am not trying to send my friend a new .exe everytime I make a change to a game mechanic - since that's expected to happen a lot as the primary use-case for the thing. If I add assets, then yeah - although ive seen games have the ability to download assets when connecting to the server, so maybe not even that. Ideally, I'll have a fixed set of assets that cover everything of interest, and the way they are rendered and the multiplayer networking and all will be consistent throughout.

hopefully this all makes sense

To answer your question more directly, we have not seen the engineering, design, and marketing skills required to do this widely available in the market, or open source space. We know of a small handful of people in the world who have the combined knowledge to do it, but none of them are working on such a product except for some of the engineers at Facepunch and Planimeter.

One person can absolutely do it, but statistically we haven't seen it.

It requires you to primarily have an exceptional background in general software development, game networking, embedding scripting languages, user-interface design, and marketing (if you actually want people to use your product/play with your software).

The guys that I know of who have the skillset to do this are all working for studios.

But the idea is the same. We see that there is a market for developers who want to play with game software, they don't want to publish a full blown game. We do want to provide a hatch for them for when they want to graduate to creating a standalone product, and that's what Planimeter Game Engine 2D is for.

So we provide abstractions like game modes, and the ability to send game mode code over to people connecting to a server who don't have it installed.

And so that developers can work on the game mode as it is being played with friends, the game engine includes live reloadable assets and real-time scripting.

Is this feasible for one person? I mostly built the game software over a decade with other volunteers from Planimeter.

I suspect we don't have many products like this in this space because most game developers aren't successful in simultaneously creating their products with a very focused scope and marketing them through discussions like we're having here now. It's literally just a constant function of engineering labor and time spent on sites looking to help people out.

It makes sense. It's what we've been trying to do for years. There are basically no products on the market that do what you and I are looking for, and my team has slowly been trying to build it up.

We have some internal projects where we have spent time looking into other game engines besides our own to see if we could move faster on established products while minimizing risk, but the landscape is amazingly poor.

For example, Unreal Engine doesn't have a fixed timestep, and its entire networking model since the late 90s, early-2000s was one based on the premise that connected clients are approximations. This is a hard departure from, say, the Quake-family of engines or engines that are designed similarly.

Believe it or not, for reasons like this, Unity is probably more appropriate for games like first-person shooters as a result. Larger studios have to jump through significant hoops to get Unreal to behave properly in common FPS scenarios, which is a technical reason why players report poor "netcode" in Unreal Engine games. You can't create reliable replicable states in UE 4 or 5.

While Unreal has some rudimentary networking out of the box, as far as we understand, Unity has issues with not providing multiplayer out of the box essentially at all? It's been some time since I've looked into it. Anyway, we expected more from the larger two institutions.

Godot also doesn't provide multiplayer out of the box, you have to write it all yourself with some bare bones abstractions you are provided, but they barely built on enet itself.

So ideally you and I would have access to a 2D or 3D engine that had multiplayer out of the box, could write game mode code that was networked to connecting clients even if they didn't have the game mode, and allow them to download assets they were missing as well. And while you were playing the game, had the ability to change the game code on the fly and send it to clients so you could develop and play test in real-time.

It doesn't exist as far as we know. Garry's Mod, Half-Life 2: Sandbox, and Planimeter Game Engine 2D are the only products I'm aware of that allow you to do this, and two of them are Source Engine games/mods.

I'm surprised it's considered a niche, but a lot of people seem to like the game engine with UI design thing over game software with robust multiplayer and live reloading.

Worst yet, a lot of amateur game developers don't seem to think multiplayer is important, or they think games vary widely in networking implementation. They really don't.

I hope more game software comes out with "game sandbox"-like features, because we didn't want to be the ones building it 11-12 years ago, but Planimeter and Facepunch seem to be the only orgs doing it.

If you have found some other software like this, though, I'd love to hear about it.

i hadnt done a real deep dive yet, but from introductory readings on a few engines this was my impression too on what we are looking for in general. Glad to know im not the only one to feel it.

>So ideally you and I would have access to a 2D or 3D engine that had multiplayer out of the box, could write game mode code that was networked to connecting clients even if they didn't have the game mode, and allow them to download assets they were missing as well. And while you were playing the game, had the ability to change the game code on the fly and send it to clients so you could develop and play test in real-time.

Yeah pretty much, though the last part would be icing - restarting the game / reconnecting to the server in response to anything more than parameter tweaks is okay. just really want to avoid excessive manual file management

>If you have found some other software like this, though, I'd love to hear about it.

I wish. If there are, then they dont support 3D PC games, because my search has been limited to that.

given how long ago garysmod was, i would be very surprised if no one has tackled those principles from the game engine standpoint already. I'm guessing any team that prioritized multiplayer sandboxing in building the game engine must have made the engine propriety and built a game studio around it.

Is Planimeter Game Sandbox something I can check out? or is that more of a target and Planimeter Game Engine 2D is the current iteration?

My next research step for myself was going to be to look at the tools in the WebAssembly ecosystem

Edit: It's probably just easier to mark our private repository as public for today. It contains copyrighted material as well as some assets from commercial games we use for internal testing so I'm essentially leaking an internal codebase, but we just don't have enough eyeballs on this for it to matter. I'll mark it private by the end of the day.

Keep in mind the main branch is probably broken from our experiments with uplifting the UI to use Yoga. CSS 2.1 and Flexbox are on the product backlog, but I may end up ditching Yoga since it's becoming too much of an engineering time drain. Feel free to git bisect it until you find the working commit. My apologies.

That being said, the Planimeter/game-engine-2d repository is only a few major features behind this branch.

https://github.com/Planimeter/gridsandbox

I am seeing similar trends to you wrt Generalists, and this is annoying but good advice.

Even when hired, generalists tend to get stuck with all the most tedious work from every department that simply no one wants to do. You end up being a one-stop shop for all intern and junior related work, with the added bonus of being invited to a lot of meetings and not getting credit (appropriate $ compensation) for your influence in those meetings.

I've settled on the belief that general knowledge is something you do for yourself, and to help set you apart from other specialists. It doesn't pay you directly. I've got almost a decade of general business and analytics knowledge, and always seems to stall out on the climb in ways you describe (not dropping the right software name brand). Time to make a bet

someone has to do it, and if the pay is adequate, it's a matter of attitude.

not getting credit (appropriate $ compensation) for your influence

i don't understand this part. you already have your salary, as negotiated. are you saying that such a positions are always underpaid?

and if such a position really has influence, it actually sounds appealing to me.

In FAANG your salary is only a small part of your compensation.

At end of year, based on your perceived performance, you get:

- a bonus (10 to 20+% of base salary)

- a stock refresh (80 to 120k over 4 years for instance for entry level/mid-level)

- a merit increase of your base salary

All these 3 influence your compensation significantly.

You can do the math, someone starting at let's say 140k base with 4 years of 15% bonus, 80k refresh, 2% merit raise, will have a different compensation than someone who's also starting at 140k but with 4 years of 20% bonus, 110k refresh, 4% merit raise

Generalism is more difficult than specialism to leverage in that initial notification. And if you end up doing work that's spread across multiple departments then it's generally harder to advance within that organisation.

>someone has to do it, and if the pay is adequate, it's a matter of attitude.

the pay is adequate for the value of the work (which is low), and it eats up your time which actually lowers the average value of your work to the company. it isnt anything other people cant do, it's just the stuff that piles up because it isnt really critical to get done but should still get done. think about what you would have a coworker help you with at your job if you had a lot to do and they asked you if you needed help with anything.

>are you saying that such a positions are always underpaid?

Yes they are underpaid because the job responsibilities are usually pretty fuzzy. you are typically given some basic responsibilities but then expected to find more work to do yourself via talking to people. so its on you to both find valuable work to perform using your general knowledge (kind of fun / interesting tbh) and also somehow be convincing that your contributions are better than what they would have gotten from an average generalist (nigh impossible). It becomes very hard to get people to recognize you going above and beyond, which is necessary for raises. So you are very dependent on having an incredibly observant manager who applies above average attention to detail when reviewing you.

how you get this work as a generalist, btw, is you ask people what they need help with. It isnt an issue with attitude, it is just one of relevance. An average generalist can be fine with this, but if you are above average at multiple roles then it becomes a point of opportunity cost. you will never work on the high-value things that someone else more specialized at the company is capable of working on. you will work on the things that were preventing that person from spending more time on the high-value things. they might talk to you about it and you might give them thoughts on the work, but it will be so casual as to be awkward for them to give you any credit for it.

>if such a position really has influence, it actually sounds appealing to me

It can be a very enjoyable position BUT you stall out - it will typically fall under some sort of generic business analyst job title at a small to mid-size company (ie not at a company with an analytics department for you to advance in). youll be encouraged to "build" that department by yourself, in your downtime, without any approved budget for it, without adding anything to anyone elses processes (ie requiring them to stick to a data entry format). So people at the company will typically like you, but you'll hit a lot of resistance trying to get past like $70k (near NYC). It makes me think that 1 good generalist is valued close to but beneath 2 junior employees with a bit of different specialization each.

the only way I can recommend a generalist position is if you are buying significant amount of stock in the company, fully believe in the product/service, and understand that sometimes in order for a team to do its best there needs to be a thankless support player somewhere in there.

you dont have noticeable influence, btw, you just know what you did and feel personally good about it. you get to sit in on meetings, typically as a note-taker (because you offered to and it makes sense because you have a bunch of misc responsibilities anyway), which means you get to make sure the most important things from a meeting are emphasized, questionable things are highlighted, and you can speak up in the meeting itself to help address misalignment's before they happen. the meetings themselves tend to get credit for your contribution rather than you, albeit sometimes you can make pretty direct call-outs that will get you credit for.

the general vibe was kind of like, you are an alert system and garbage collector. a lot of the time the alerts are received as helpful reminders of issues that would have been caught somehow anyway. but that's kind of a catch-22 for proving your value. you feel fairly confident they wouldnt have caught the issue based on the nature of the due diligence you applied to the situation and your intimate knowledge of the related business operations. you cant exactly point why you think the issue would have gone unnoticed without throwing someone under the bus, which isnt fair to do before the issue actually arises. the catch-22 is that the issue isnt going to arise because you pointed it out, but you cant prove it was necessary for you to point it out unless you dont point it out.

So you need an observant, reasonably skeptical manager on the same page as you to notice the shitstorms you prevent. But even then there's typically no direct means of compensating you for it. you get labelled a good employee, quickly jump to that 70k area pay cap into yearly inflation adjustments, then you end up just getting more soft benefits like openness to alternative work hours, no resistance to taking time off, and such. With the caveat that you have to correctly read the room on these things and assert them yourself because you wont be getting an email detailing such perks. that is all well and good, but it seems better fit for someone near retirement then someone with goals of trying to buy a house and raise a family.

i was thinking people who are probably capable of retiring, but still want to work a job with more soft perks than pay. but yeah, maybe the work doesnt really suit anyone. Just a temporary type of work until you figure out what kind of specialization you want to do, or to build up experience for consulting maybe

I think if such practices became mainstream, then people would begin to realize the difficulties involved in having a coherent conversation about a point of disagreement. The passive-aggressive cudgels you mention would fall flat because it is instantly noticeable as not fitting the necessary patterns for coherent conversation.

No one needs to accept or reject the other. the point of an argument is to educate each other, answer questions, and allow each other to fit both opinions into their own world view (ex. what are the limits / specifics to your belief?). Maybe someone changes their mind during that conversation, maybe they dont. maybe someone needs to let the new information ferment in their minds and life for a bit before it clicks. that is part of giving someone benefit of the doubt

maybe you are talking to someone who is venting, or in the middle of a mental episode (i mean that literally, not derogatorily), or was unfortunately born a narcissistic manipulator and cant help themselves. It doesnt matter, they cannot "win" the conversation - no one can - and the more people that realize this truth about conversations in general, the better the world will become

This is why I stopped giving people benefit of the doubt. With every uncharitable action of theirs, I tried to understand their perspective, explain mine, and discover a ground truth. Yet, they are not interested in finding the truth. They are interested in doing what they believe is to be true, regardless of whether it is true or not. They are not interested in getting educated. They have already decided that they are educated and I am wrong, just because we have different ideas. My strategy of giving benefit of the doubt in such a case turns out to be nothing but a waste of time.

Therefore, I've changed my strategy. If they don't respond well to my giving benefit of the doubt, I'll confront them directly. If they insist, then they'll become an out-group to me. Starmanning no more.

I dont enjoy casting people into out-groups, but I understand its appeal and necessity for some people. However, I dont give BotD for the other persons benefit. I do it for my own peace of mind and because I find it yields better conversations overall. Sometimes conversations go nowhere, or people deliberately try to manipulate. Those conversations yield nothing, but that's okay. No one owes me anything, afterall. there's always the future.

I find the best way to keep people open to changing their minds about something either now or in the future is to make them feel like they are free to support whatever they want. Then I act as a source of information and act as a safe zone for thought-exploration.

People do not enjoy feeling hunted. and that goes for educated people hunting uneducated people in order to teach them something or else be out-casted. I think a response of "fuck you, go ahead and outcast me" to that would be pretty normal.

I would much prefer in-grouping people with proper differentiation. "so and so is a great cook!" instead of "so and so doesnt understand climate change and couldnt hold a rational conversation with me about it that one time so I dont associate with them anymore and if they die then good riddance". People want to be validated. Sometimes finding something to validate someone on rather than attack them on the points of disagreement can help reshape their identity. they might never agree with you, but maybe you can move them towards ignoring the topic entirely within their lives in favor of other things that are better aligned with the good of society.

The types I'm talking about are the ones who are not willing to cooperate. I act as a team player, yet they have their own agenda against the team's. What's worse is they act as if they are willing to cooperate, to benefit from BotD. I don't know whether they do so consciously to manipulate people or because of their insecurities. It doesn't matter. If they are not willing to cooperate after being treated with the best of intentions, then those good intentions are better invested to where they are valued.

It reminds me of a psychology study done on kids that I wish I had the source for. IIRC when asked to perform a task for no reward (like, how many simple math problems can you solve in a minute?), many performed it well. then they gave the good performers gold stars. Immediately, the kids not getting stars started performing worse than originally. Then they took away the gold stars. Now the top performers started doing worse as well.

People just want to be seen and heard. Measure the things that are important WITHOUT deciding in advance what is considered a "good job". this way you can do your 1:1 on the whole picture. "you're slow, but your quality is exceptional"

The only exception to this is if the work is simple enough that you can measure it directly against your $$$, and you have a target that ensures you are green. Like maybe a strictly data entry position for a very consistent source of data. In this case, there arent 1:1's to be expected with your employees.. they either hit target or get fired.

Metric is closed tickets? OK, open more smaller tickets, close aggressively and tell customer to open a new ticket if problem continues.

You just have to look at public education being measured by exam scores to see the effects of this writ large.

Creating a new metric and seeing it go up for a year or two is good. Creating a new metric and seeing it go up and up and up past that ought to be considered suspicious. Not automatically "wrong", but something to dig into, rather than celebrate on cognitive autopilot.

For example, getting from one nine of uptime to two nines isn't all that hard or expensive (a month of downtime per year to half a week). Getting from three nines to four nines is a lot harder, and a lot more expensive, and if you tie success at the organization to that improvement, people will fudge to reach it.

Nearly every metric is like that. Amount of sales, tail latency, recruiting efficacy.

SPC is tricky case in that regard. Because it, kind of, makes the trend and long term development of metric the target.

Edit: You had me thinking. Settng a target, or rather target tolerances, for a metric can be necessary if that metric is directly measuring the output of one process that is used by another process. E.g. forecast accuracy, if downstream ops are expected to cope with, say, a deviation of 20 %, that tolerance kind of becomes the goal for the forecast accuracy. This is not some arbritary goal so.

For me, setting limits on certain metrics is something different from a target. It is a thin line between those.

I think a forecasting target of 20% shouldnt be necessary unless you want to be around 19% deviation. Which might be the case if trying to reach 0% has diminishing returns and isnt actually cost effective. In other words, targets make sense when you are literally targeting a sweet-spot.

If 0% deviation is desired, then the directive is still the same - try to be as accurate as possible. If "as accurate as possible" is consistently falling below 20% (which is the importance of measuring and data) then you need to address that. Maybe forecasting is broken or infeasible.

On ops side, they might not be expected to handle deviations over 20%, but the directive is probably the same either way "do the best you can to manage the deviation".

>For me, setting limits on certain metrics is something different from a target. It is a thin line between those.

I agree with this in general. It can be tricky to communicate "here is a boundary that absolutely cannot be crossed, but keep in mind you should be nowhere near that boundary anyway" without it impacting peoples personal optimization strategies. You really have to weigh the consequences of what actually happens when that limit is reached, and if it's worth a possible depression of performance from the team.

It all comes down to leadership, doesn't it?

There are a lot of them with similar results, going back to 1960s.

Kohn "Punished by Rewards" is a good general start for that stuff. I think there was some generalised study/review published in late 90s. Probably Deci and Ryan (?).