ZeroCool2u

m6a.48xlarge: 192 vCPU & 768 GiB

If that's not a large node for you, well you're living in a different world from me. Yes to using EBS.

I've done a good amount of geospatial analysis for work.

One thing you quickly realize with geospatial data is that it's incredibly nuanced. You have to be quite careful about understanding which coordinate reference system (CRS) and for visualization which projection is being used. The CRS is somewhat paranoia inducing if you don't have great infrastructure setup with the right tools to carry that metadata with your geospatial data.

I've tested everything AWS has to offer, Postgres/PostGIS, Spark/DB, Snowflake, Trini, and ARCGis.

I'm convinced the best tool for large scale geospatial work is Google BigQuery and it's not even close. It took an expensive multi hour query running on PostGIS deployed on an enormous m6a EC2 instance to less than 5 seconds that ran in the BigQuery free tier. It does make sense if you think about it, Google was early with Maps, but it is just stunning how much better they are in this specific niche domain.

This was using publicly available FEMA data that Snowflake and AWS services would just choke on, because the geometry column exceeded their maximum byte size. Spark doesn't even have geospatial data types and the open source packages that add support leave a lot to be desired.

This guy is running on-prem, so maybe this made sense, but I just would never bother. The storage for BQ would probably be less than $100/months for 20 TB.

In the linked article it says:

"Unlike AWS Aurora we decided to open source all the changes in Postgres and also send them upstream as well as fully open source our cloud native storage."

Also, at the top right there's a link to the GitHub repo: https://github.com/neondatabase/neon

You can find replies from folks in this thread as well discussing self hosted deployments and you can find an entire discord channel dedicated to that topic here: https://discord.gg/92vNTzKDGp

I'm not associated with Neon and I've never used it before and I certainly don't disagree with your sentiment, but it seems like Neon has gone above and beyond in making sure that workloads could continue to run in the event the company fails and broadly assuaging those concerns.

"- Python's dict is a fast unordered hashmap. However, if you need order-aware operations like C++'s std::map::lower_bound(), you're out of luck; Python's standard library doesn't have a tree implementation."

I'm fairly certain that in Python 3.7 and later standard library dictionaries are now ordered by default.

Which seems like their biggest mistake. If they would just release a card with more than 24GB VRAM, people would be clamoring for their cards, even if they were marginally slower. It's the same reason that 3090's are still in high demand compared to the 4090's.

Personally, I've never really been comfortable with the concept of stock buy backs. We already have dividends and obviously the redirection to R&D for increasing shareholder value. It doesn't seem like something that's healthy for the market as it's as close as you can get to legal price manipulation and obfuscates the "true" price.

Does anyone have a market theory justification that makes them comfortable with the idea?

Fly looks great to me, though I've never had a chance to use it. For what it's worth though, Cloud Run on GCP is one of my top 3 favorite infrastructure/deployment tools, so you're setting the bar pretty high.

Thank you for the answer! Very interesting.

That is metal af ;)

SIMD can be pretty architecture specific. For example, does the CPU support AVX-512 or SSE3? So, you have to have a few code paths if you're going to support a wide variety of hardware.

I don't have an answer to your question it just occurs to me that maybe the Linux kernel doesn't allow a lot of SIMD for this reason or maybe they require a fall back/slow code path to be available if you're submitting a patch that includes SIMD operations?

Curious if anyone knows the answer to this.

Hosting by Mistral/OpenAI/Startup is often a non-starter for the larger enterprise style customers.

For example, they have a legal agreement with Azure/GCP/AWS already and if they can say it's "just another Cloud provider service" it's stupid how much easier that makes things. Plus, you get stuff like FEDRAMP Moderate just for having your request sent to Azure/GCP/AWS instead? Enormous value.

Getting any service, but especially a startup and one that ingests arbitrary information, to be FEDRAMP certified is the bureaucratic equivalent of inhaling a candy bar.

We have the Unlimited Plus plan and that comes with Google One, so it seems there's some accounting logic they haven't figured out how to reconcile.

I'm a Google Fi subscriber, so I still can't switch to the Google One plan with access to it. Incredibly frustrating.

So uv uses pubgrub-rs, a rust implementation of the pubgrub version solving algorithm originally written for the dart language. I suppose I shouldn't be surprised, but I always loved Pub the Dart/Flutter packaging tool. It's the only thing out there that comes close to cargo that I know of. Fun to see these chains of inspiration reach across languages.

That's what they attempt to cover, but they're not super successful at it.

This seems nice and the plots look great, but I have a hard time imagining switching to Observable from Plotly since there doesn't seem to be a way to make any plots interactive. By which I mean Zoom and Pan. The nearest point high light feature is nice, but what if I want to zoom in? None of the examples here seem to be able to do that and quick google search doesn't make it seem like that's straight forward. That's not even additional code when I use Plotly, it's just built-in.

There's also the issue of convincing staff to use JS instead of Python which is still just a tough sell. I think everyone on my team (of data scientists) would look at me like I've got two heads if I were to suggest that. Maybe we're not the target demographic though.

I do like the idea of shipping the data straight to the client, but I don't have a lot of confidence in our corporate network doing well and not slowing stuff down. Perhaps the graphics all are sent pre-rendered over the wire though? I'm not sure, but would be cool if Observable figured out a way to side step that issue.

I suppose you could try with a Google Colab notebook attached to a free TPU instance? Probably would be quite limited if it worked at all.

I think last year or so there was an issue where Microsoft Teams broke the phones ability to dial 911 in certain situations. It was something about Teams taking over the dialer or something? I can't recall exactly. Google understandably took a huge amount of flak from users for that and the FCC got involved. It was fixed, but the reputational hit was significant and I'd bet they're a lot more aggressive with any apps that interact with the phone/dialer in general now. Especially ones that aren't being consistently updated.

If I combined ParadeDB with a common postgres extension like PostGIS, would I get the benefits of both?

This is the content of the CLA:

Zed Contributor License and Feedback Agreement Welcome to our Contributor License and Feedback Agreement! Here's a quick breakdown of what's inside:

You (the contributor) are entering into an agreement with Zed Industries, Inc. You're giving Zed permission to use and share your contributions (like original works or modifications). You assure us that the contributions are truly your own and you have the legal right to share them. You're not required to support your contributions but you're welcome to if you wish. If you ever notice an error or change in the details you've given us, you agree to let us know. In short, this agreement covers the terms for your valuable contributions to Zed's projects.

Legally binding section follows: By submitting a Contribution, use of the Solution or any components related thereto (as such is further defined in Zed’s End User License Agreement located here: https://zed.dev/eula) or as enabled pursuant to your use of Zed’s collaboration tools offered therein, you hereby accept and agree to the terms and conditions set forth in this Contributor License and Feedback Agreement (the “Agreement”) for Your present and future Contributions submitted to Zed Industries, Inc. (“Company”). Except for the license granted herein to Company and recipients of software distributed or made available by Company, You reserve all right, title, and interest in and to Your Contributions.

1. Definitions.

“Contributor”, "You", and "Your" shall mean the copyright owner or legal entity authorized by the copyright owner that is making this Agreement with Company. For legal entities, the entity making a Contribution and all other entities that control, are controlled by, or are under common control with that entity are considered to be a single Contributor. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity.

"Contribution" shall mean Feedback (as defined below), any original work of authorship, and any modifications or additions to an existing work, that is intentionally submitted by You to Company for inclusion in, or documentation of, any of the products owned or managed by Company (the "Work"). For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to Company or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking or collaboration systems that are managed by, or on behalf of, Company for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by You as "Not a Contribution."

“Feedback” means suggestions, comments, improvements, software code / snippets or other information submitted to, shared with or otherwise made available to Zed or its contributors with respect to, or in connection with the use or interaction with, the Work, Zed Network Based Service or Editor technology (as defined by Zed and as further described within Zed’s End User License Agreement located at https://zed.dev/eula).

2. Grant of Copyright License. Subject to the terms and conditions of this Agreement, You hereby grant to Company, and to recipients of software distributed by Company related hereto, a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare derivative works of, publicly display, publicly perform, sublicense, and distribute, Your Contributions and such derivative works (the “Contributor License Grant”). Further, to the extent that You participate in any livestream or other collaborative feedback generating session offered by Company, you hereby consent to use of any content shared by you in connection therewith in accordance with the foregoing Contributor License Grant.

3. Grant of Patent License. Subject to the terms and conditions of this Agreement, You hereby grant to Company and to recipients of software distributed by Company a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by You that are necessarily infringed by Your Contribution(s) alone or by combination of Your Contribution(s) with the Work to which such Contribution(s) was submitted. If any entity institutes patent litigation against You or any other entity (including a cross-claim or counterclaim in a lawsuit) alleging that your Contribution, or the Work to which you have contributed, constitutes direct or contributory patent infringement, then any patent licenses granted to that entity under this Agreement for that Contribution or Work shall terminate as of the date such litigation is filed.

4. You represent that you are legally entitled to grant the above licenses, including but not limited to any video content shared or recorded as related to collaboration with Zed and the Zed community. If your employer(s) has rights to intellectual property that you create that includes your Contributions, you represent that you have received permission to make Contributions on behalf of that employer, that your employer has waived such rights for your Contributions to Company, or that your employer has executed a separate Corporate CLA with Company.

5. You represent that each of Your Contributions is Your original creation (see section 7 for submissions on behalf of others). You represent that Your Contribution submissions include complete details of any third-party license or other restriction (including, but not limited to, related patents and trademarks) of which you are personally aware and which are associated with any part of Your Contributions.

6. You are not expected to provide support for Your Contributions, except to the extent You desire to provide support. You may provide support for free, for a fee, or not at all. Unless required by applicable law or agreed to in writing, You provide Your Contributions on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE.

7. Should You wish to submit work that is not Your original creation, You may submit it to Company separately from any Contribution, identifying the complete details of its source (e.g., attribution) and of any license or other restriction (including, but not limited to, related patents, trademarks, and license agreements) of which you are personally aware, and conspicuously marking the work prior to submitting it to Zed or the Zed community.

8. You agree to notify Company of any facts or circumstances of which you become aware that would make these representations inaccurate in any respect.

Gift article link here: https://www.wsj.com/arts-culture/television/jon-stewart-retu...

Tbf JAX is super nice. Even easier than PyTorch in many ways and astonishingly fast. XLA is super powerful.

We're watching Apple fill the moat in.

Wow, this is _so_ much better!

I wish Skydio still did reasonably priced consumer drones with these features. Their old demo videos were glorious.

Great catch!

I'm sure they've done benchmarking comparisons with the A100 and H100, but since they're not sharing them it seems unlikely they show the TPU's in a favorable light. Plus, they pick up more spend every time someone runs these benchmarks for themselves.

It seems to be. Bard is only using the G-Pro model, not the Ultra, which is what all the benchmarks they're touting are showing. If I had to guess, the best you could hope for is exactly what you're describing.

The performance results here are interesting. G-Ultra seems to meet or exceed GPT4V on all text benchmark tasks with the exception of Hellaswag where there's a significant lag, 87.8% vs 95.3%, respectively.

Much more interesting link: https://deepmind.google/technologies/gemini/