One possible explanation: business owners have more skin in the game and care the most, so they are the most demanding and can’t tolerate waste. So they are the hardest to satisfy.
Additionally, they are not used to mincing their words because they don’t have bosses and are the most direct (and also egoistic).
I’ve owned some trivially small businesses, and something I think is missing in the discussion is how quickly people’s presumptuous bullshit has me on the phone burning time-money. They think I’m rude, it’s them though, and their sloppy sales ‘lies’.
In those roles it’s not that I’m king whatever of wherever, but I know intimately how much my hours capitalize for, and given that lots of sales/customer communication is happening on the phone, you get a big impedance. Smug low tier sellers trying to talk past me, and flatly disrespecting me when trying to rationalize their pre-packaged pitch.
Architecture is customer facing. I am nice, and soft spoken the first time. But when kindness is misunderstood as ineptitude and people start talking to me like I’m a girl who could just never, like, understand a transmission…? My only regrets are not hanging up hard enough and refraining from swearing at them in the early years.
Other people’s C-tier employees and D-tier operations are not my problem unless and until they pay me.
According to NYT it seems like there were 2 controllers and “2 more in the building”. They also wrote that 2 seems normal for the late slower time of the night.
Not saying this is the right number of controllers to have, just sharing what I read in NYT.
I know about 2 big discontinuities in group dynamics, which are based on the limits of human cognition at specific sizes:
1) ~7 people. This is when each member cannot participate in the whole context and all important decisions. "2 pizza teams", limits of the working memory. Decisions cannot be done all together anymore. This results in hierarchies forming and people worrying about their positions in them.
2) ~150 people (Dunbar's number). This is when group members cannot all know each other anymore and have meaningful relationships. Max sizes of family-based tribes, an important unit size in the army. This results in inability to observe each other's actions well, so understanding who contributes a lot vs not shifts to indirect stories. Group members start building narratives instead of demonstrating contributions directly.
Beyond the common narrative on this topic, a factor to consider is that people might be more interested in hearing about death causes, which are not considered their own “fault”. These situations are less “fair”. Thus terrorism, homicide and accidents get a big focus.
Why do you think Glean and other top company GPT startups who are doing this for longer and have more resources cannot get this level of performance (helpful and accurate)? What makes your approach different and not easy to replicate?
Great question. We are actually surprised more than anything that existing products aren't better. Other companies' algo work DID get disrupted by reasoning models, because newer models don't care about what the top 3 search results are, they're really great at reading hundreds of documents and picking up signal from noise.
I can't speculate exactly on the work that other companies have done, but my guess is we were way more focused on the type of questions people actually ask to each other. We know there's still a lot more you can do to continue to improve it, and so it's up to other folks if they do it too.
Lastly, a few ways we want to be distinguished from all the other offerings are: 1) super easy to setup, 2.) very developer friendly
It goes in gram increments and my laptop was able to read 7300g pressing as hard as i could, which I was surprised it would be designed to read that high, might go up to 10kg but I don't want to crack my trackpad lol. The actual measurements though are extremely unreliable. I've found it can't reliably measure anything, measuring a roll of tape gave me measurements from 70g to 700g, it always settled on a number but was different every time. Maybe the underlying data is more accurate but this API is definitely just designed for outputting the force of a finger. M1 MBP for reference
For starters, according to that link it only left "experimental status" in Oct 2023; it's pretty new. Although their GitHub still describes it as "an experimental language". That doesn't exactly inspire confidence for long-term production use.
The second problem is that while imperfect, everyone and their dog knows at least the SQL basics. There's a lot of value in that.
Thirdly it's written in TypeScript, so anyone not using TypeScript will either have to run some sort of "generate" step, or have to rewrite it in $language_of_choice. Both are painful. This is not a TypeScript problem: you will have that with any language (although with languages that compile to a binary it's a bit less painful, albeit still painful).
Lastly, it doesn't work for all SQL flavours: just BigQuery and PostgreSQL. That's pretty limited.
All of that is assuming the SQL it generates performs as well as "native" SQL, and that all of this can be reasonably debugged if something goes wrong.
I see Malloy as an analysis tool like Pandas, PowerBI, Tableau, or Looker. It seems like most people write SQL to get "All the data" and then take it to an analysis tool for further study.
With Malloy you can do the analysis on your data lake directly for 80% of your questions.
Languages like Malloy or PRQL require some upfront investment to learn and setup. Additionally, I think both Malloy and PRQL are query oriented, so you still need to learn a decent amount of SQL to interact with your database (for updates).
I'm guessing most people would rather spend effort in solving their immediate problems with SQL rather than bet on a newer and less known technology, even if has a lot of promise.
Which is one of those cases where I briefly want to reject linguistic descriptivism because to me the "G" in "AGI" is precisely "general".
But then I laugh at myself, because words shift and you have to roll with the changes.
But do be aware that this shift of meanings is not universally acknowledged let alone accepted — there's at least half a dozen different meanings to the term "AGI", except one of them requires "consciousness" and there's loads of different meanings of that too.
Additionally, they are not used to mincing their words because they don’t have bosses and are the most direct (and also egoistic).