How do you think it should be treated? I think at the individual granular data point level adding a tag or note about the equity not being immediately liquid is a good start. But I don't think it'd be a good idea to weigh the stock differently since that can depend on so many things. For example SpaceX and some other private companies do offer regular liquidity and I would consider their equity close to liquid.
Appreciate the feedback though, and definitely agree we can work on how we display the data and make it more clear.
1. Salary (straightforward, on regular schedule, and you'll get it)
2. Bonuses and RSUs (various vesting rules, and ways you can never see it)
3. Startup stock and (worse) stock options (probably worthless, vesting rules, and you might need an advisor to make sure you don't exercise and come out with a big negative)
This is a great point, and something we plan to address. We currently use Nielsen's DMA (Designated Market Area) mappings within the US to separate out regional areas which was used for TV / media market surveys. We happen to use DMA categories for our regional pages on Levels.fyi which is why it was easiest to start with since we already had this data captured. The features can sometimes be a bit off and seem like they're grouped very far and wide (you'll notice there's a bit of Denver within Nevada and its just a vestige of how it used to be categorized), but it still provides a bit of a broader level grouping than something like zip code. We've also been considering using Combined Statistical Areas using population instead, but the benefit with DMAs is that it offers full coverage of the entire US whereas some major tech hubs are still missing from CSAs if relying solely on population.
We're planning to create some of our own regional definitions and borders using our own submissions and that should offer some more tighter bounds. This was just a v1, and I think its already resonating with folks.
> We've also been considering using Combined Statistical Areas using population instead, but the benefit with DMAs is that it offers full coverage of the entire US whereas some major tech hubs are still missing from CSAs if relying solely on population.
I think just using the 387 MSAs [1] instead of the 181 CSAs would get you far enough to cover all the major tech hubs.
If that data is submitted by individuals to a particular company, is it possible to see a lot more detailed heatmap, perhaps down to each address of each company?
I think that part is still a bit ambiguous, it's almost how people and companies self-identify the role. But that said, we collect a bit of a taxonomy for our role structure, and we specifically look at Software Engineers focused on AI. The responsibilities can still differ from company to company, but that's what we used for our dataset.
Congrats on the launch! Awesome to see this release as an early user who was able to check it out. The shared workspaces and shared browser windows with context in place has been incredible for collaborating with folks. We have our Figma design, Notion doc, and Gitlab MR all in the same space so we don't have to go searching for each one independently or have them cross-linked to each other.
Another framing of this is whether companies treat their engineering organizations as a cost center or a profit center. Cost centers suppress salaries as much as possible optimizing the budget. There's little growth within these companies which is why the zero-sum philosophy is passed down to their compensation strategy. Whereas profit centers encourage more and more investment as compounding profits come in from previous dividends. Growth is what powers everything, higher pay has a positive-sum gravitational pull on talent sustaining a flywheel for more profits to come in (hopefully).
All the companies that pay very well such as on https://levels.fyi/2023/ are profit centers encouraging investment in talent, competing for the best across companies because they know its worth it. Each hire even at extremely competitive wages will make back their salary manyfold if they're successful.
That’s fair, but 0 guardrails just entirely minimizes the incentive to create a custom GPT.
I was also thinking about when there’s paid access to certain GPTs. Someone can just download the files and spin up their own? Doesn’t seem valuable for devs. If that’s how the incentives were intended then sure. I just don’t expect many people to be enabling use cases with it.
I think the GPTs market is a bit of a novelty thing. But it might work with low costs. I've been making my own GPTs last few days and I find the capability useful but very inconsistent.
If you give a GPTs personality more complex tasks, it gets distracted and reverts to its default ChatGPT personality.
There aren't enough parameters and compute in these LLMs for the things we're pushing them to do. We need hardware to catch up and offer analog compute for neural nets, so we can scale them way up. And of course evolve architectures to make use of that capability. Only then I see this becoming truly rock solid.
Also using the context window for personalities is likely the wrong approach. You gotta finetune this on the model itself.
But yeah GPT can't keep a secret. But humans also can't, so I guess it's imitating us fine.
Most companies still don’t necessarily list total compensation. For base salary though, these ranges are quite helpful and has definitely caused some “good” trouble increasing baseline wages among competing companies.
At higher levels it’s still ambiguous what ranges truly are since job postings can span multiple job levels (which is why some ranges can feel dubiously wide). Ended up writing a bit about this phenomenon here and what contributes to it: https://www.levels.fyi/blog/notes-on-california-transparency...
Appreciate the feedback though, and definitely agree we can work on how we display the data and make it more clear.