Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
m_ke
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
39 ms
·
511.
▲
by
m_ke
5y ago
They could be off by an order of magnitude and it would still be pretty bad. Here are their official numbers from Q2: https://s21.q4cdn.com/399680738/files/doc_financials/2021/q2...
512.
▲
by
m_ke
5y ago
Because their products are used by over 2 billion people a day and according to their own research it's having a negative impact on a significant chunk of them.
513.
▲
by
m_ke
5y ago
Growth at all cost: https://twitter.com/JesseLehrich/status/1443451725768105986
514.
▲
by
m_ke
5y ago
Deep learning is narrowly applicable to every domain, that's the beauty of it. It's delivering 1.1-10x efficiency improvements for a lot of common workflows, which might not seem that impressive but really adds up. My cousin is an
515.
▲
by
m_ke
5y ago
Yes we use GCP for production workloads and enjoy the benefits of being to scale at will. I'm strictly speaking about hosting large multi TB datasets and running machine learning training jobs that end up costing thousands of dollars e
516.
▲
by
m_ke
5y ago
My original comment was referring to companies who train their own machine learning models. They might not be spending the way large slow corporations and the government do but there's a lot of investment in the space and a ton of room
517.
▲
by
m_ke
5y ago
No sane person is sitting on mountains of cash.
518.
▲
by
m_ke
5y ago
Sorry to break it to you but most of HN is not well informed. Look for some threads in your area of expertise for proof of that. Only people selling AGI are marketers and futurists.
519.
▲
by
m_ke
5y ago
Sounds like we need to start taxing the rich a lot more to pay it off.
520.
▲
by
m_ke
5y ago
Yes except for the ones that are lighting saudi money on fire by having softbank pay for it. A lot of ML startups end up buying hardware for training because they can get a GPU for what it would cost them to rent it for 2 months on GCP/
521.
▲
by
m_ke
5y ago
Cloudflare could really shake things up on the ML side of things. The egress costs and GPU prices on AWS and GPC make them a nonstarter for most companies, forcing people to rack their own hardware.
522.
▲
by
m_ke
5y ago
some good insights on depth: https://arxiv.org/abs/2106.09647
523.
▲
by
m_ke
5y ago
Most people unlock their phones using a deep learning model based facial recognition system, they talk to their devices thanks to deep learning, they translate documents with transformers, even google maps uses GNNs for ETA estimates and ro
524.
▲
by
m_ke
5y ago
No sane person believed that deep learning would get us to AGI withing the next few decades. Anyone saying so is doing it to fool VCs out of their money.
525.
▲
by
m_ke
5y ago
The recent advances in deep learning are good enough to keep industry busy for the next 20 years. Each year has brought us substantial improvements and if you look past the popular benchmarks there's no sign of things slowing down.
526.
▲
by
m_ke
5y ago
I've been saying this for a while. I used to live in queens and would take a practically empty bus down a traffic packed metropolitan ave to a coworking space in bushwick. There was no way the bus was breaking even with the 5-10 people
527.
▲
by
m_ke
5y ago
I have an image recognition service on GCP where majority of the latency is the initial upload. Pushing that closer to the user would be nice.
528.
▲
by
m_ke
5y ago
Great start but the field is moving on from tensorflow. Torchscript support would be amazing.
529.
▲
by
m_ke
5y ago
Do GPUs at a reasonable price next and you'll own the ML space. R2 will be amazing for hosting computer vision datasets and models.
530.
▲
by
m_ke
5y ago
My worst date ever was a trip to MOMA, I had to go to a museum as an assignment for an art history class. The girl must have thought I was a weirdo for taking her there and we pretty much stopped talking after that night.
531.
▲
by
m_ke
5y ago
Would love something like that for storing large image datasets for computer vision. Storing embeddings, predictions and metadata in a contiguous format with compression support, ANN indexing support and SQL would be amazing.
532.
▲
by
m_ke
5y ago
Here's the US immigration page: https://travel.state.gov/content/travel/en/us-visas/immigrat...
533.
▲
by
m_ke
5y ago
How else do you expect people to immigrate here if not on a visa?
534.
▲
by
m_ke
5y ago
I hope you're right but: 1. His approval numbers are tanking https://news.gallup.com/poll/354872/biden-approval-rating-hi... 2. Summer is just ending and we're up to 2000 Americans dying from COVID a day
535.
▲
by
m_ke
5y ago
It was not meant to be flamebait, I'm a democrat and Manchin is clearly a corrupt corporate stooge. Here's an Exxon exec confirming on tape that Manchin is his go to guy and they want the infrastructure deal killed because it woul
536.
▲
by
m_ke
5y ago
Another thing that some people don't realize is that this reconciliation bill could be the last attempt at Biden passing any meaningful legislation. If it fails dems are guaranteed to lose seats in the midterms, giving up their slim ma
537.
▲
by
m_ke
5y ago
I tried to match the tone of the headline in the article, especially the subtitle: "The Biden administration’s biggest success has not received enough attention".
538.
▲
by
m_ke
5y ago
Here's my wording for anyone who can't see it due to it being flagged: """ *Temporarily, and partially thanks to the previous administration The infrastructure bill is about to fail due to a few corrupt corporate De
539.
▲
by
m_ke
5y ago
See this comment: https://news.ycombinator.com/item?id=28647021 The $300 per child per month payments are temporary and will expire at the end of the year. Before that the extended unemployment helped a lot of parents by gi
540.
▲
by
m_ke
5y ago
It's not partisan, the child tax credit that gives parents $300/month/child that was included in the "American Rescue Plan" expires at the end of the year. [1] The new infrastructure bill was supposed to include fre
More ›