Viewing profile — mostdataisnice
mostdataisnice
HN member- Joined
- Fri, May 28, 2021, 8:07 PM UTC
- HN karma
- 111
- Public activity
- 23 items
- HN profile
- View on Hacker News ↗
About mostdataisnice
No profile information was provided.
Recent public activity
-
comment
Comment #35429552
What it comes down to is a strong, dogmatic belief at Tesla that vision will solve all problems. It's not aligned with reality unfortunately
-
comment
Comment #35350559
This is the clearest example of an attention grab I have seen - it does nothing for commercial use of Llama unless they provide a version of the weights produced by them and not Fa…
-
comment
Comment #34346019
fwiw - the lead authors on that linked paper are all grad students not employed at Databricks. That being said, they're advised by Databricks people
-
comment
Comment #33628959
As someone replied above, it's because SQL is just 1 backend and it's weird to expose an API that only works on 1 backend. Once you have many devs working together, you need a remo…
-
comment
Comment #33626827
Where does the article say that?
-
comment
Comment #33626808
SQLite is literally a backend for MLflow, so the argument being made really is that you should just use SQL when you can, which is kind of adjacent to any criticisms of MLflow
-
comment
Comment #31023836
Right - I think this is part of the "scaling" of the internet. Those of us (relatively small set of earlier heavy internet users) are forced to change as the product is built for t…
-
comment
Comment #31021372
This resonated a lot with me. Often if I'm learning a language and I see no one is doing what I'm querying for, I take a step back and try to ask a more basic question
-
comment
Comment #31021355
Another thing for us on HN to note. This is how most people ask questions . It is not just that the other way is wrong - it is also less common.
-
comment
Comment #30984971
It would be an unmitigated disaster to have this dude on the board - his decorum is totally unacceptable
-
comment
Comment #30872229
This is ridiculous - the most promising engineers are not always the best people managers.
-
comment
Comment #30633840
It's not ridiculous at all - you don't get to reap the benefits of the network without participating.
-
comment
Comment #30090419
I think what HN and the SWE community at large has just missed about Google over the last 10 years is that the product is being built for the masses. Most people would prefer if yo…
-
comment
Comment #29244624
...and there's a reason for that. Automations to update the status page are rarely acceptable, since the status page statuses have legal and financial implications. Therefore, the …
-
comment
Comment #29235338
What fake benchmarks are you talking about?
-
comment
Comment #28767214
I mean, sure? mumblemumble is still right though. If you're looking for a cynical reason for everything FB related, then, sure, it's true that a human error looks bad.
-
comment
Comment #28298388
This is just completely false - if you're working on a feature, you want a PR splash controlled by you, not a stream of silent leaks. PR begets other PR. You're being super presump…
-
comment
Comment #28291925
Delete the Waymo One app and reinstalling worked for me when I saw this bug.
-
comment
Comment #28291863
Also, I don't think we actually know what the end-of-life of Tesla vehicles is. They've not been mass produced for long enough.
-
comment
Comment #28269849
Underrated truth - have owned a Tesla for 5 years, and the number of stupid things that go wrong with it is incredibly high. It works like a phone...after a number of years, it tak…
-
comment
Comment #28092703
There's a really salient point about reliability of quality content here - I curate which podcasts I listen to, so in my downtime, I can quickly pull those up. Unless my downtime i…
-
comment
Comment #27910340
The fundamental problem is that NASA's funding is tied to politics and not a business model. Both Democrats and Republicans gutted NASA over the last 15 years and essentially kille…
-
comment
Comment #27778763
While interesting, we don't know enough about how models are learning to where we would be able to consider doing this.