ko_pivot

15 ms·

61.
by ko_pivot 2y ago
Of course. This eventuality was discussed for months prior. It’s one of the less romantic facts about Europe that goes along with the rich history; truly antiquated sanitation systems.
62.
by ko_pivot 2y ago
As much as this has the vibes of a classic OSS rug pull, as a Cockroach user, I don’t really take it that way. First of all, it was already not open source and secondly, the free to use version was missing key features like follower reads a
63.
by ko_pivot 2y ago
I don’t believe Neon supports multiple write nodes.
64.
by ko_pivot 2y ago
While you can use Supabase as simply a Postgres provider, the more interesting comparison IMO is to other backend-as-a-service providers. Supabase used to call themselves a “Firebase alternative” but at this point they have surpassed Fireba
65.
by ko_pivot 2y ago
Obviously this could be true, but also could be Musk saving face due to X streaming infra buckling under a heavy volume of legit requests. It’s happened before.
66.
by ko_pivot 2y ago
Fargate is very expensive at scale. Great for small or bursty workloads, but when you’re at Figma scale, you almost always go EC2 for cost-effectiveness.
67.
by ko_pivot 2y ago
This adds support for ephemeral EBS volumes. When a task is created a volume gets created, and when the task is destroyed, for whatever reason, the volume is destroyed too. It has no concept of task identity. If the task needs to be moved t
68.
by ko_pivot 2y ago
I’m not surprised that the first reason they state for moving off of ECS was the lack of support for stateful services. The lack of integration between EBS and ECS has always felt really strange to me, considering that AWS already built all
69.
by ko_pivot 2y ago
If this really is a strategic investment, maybe it means Citus style sharding is on the horizon?
70.
by ko_pivot 2y ago
I'm guessing AWS simply did not build reading/writing the current DB state to/from S3 into Aurora Serverless v2. Neon has done a ton of work to make scale to zero happen in terms of treating the actual database disk as nothin
71.
by ko_pivot 2y ago
As a lot of commenters are already pointing out, this is a bit different than Google’s past escapades with poor product management. In this case, they have a replacement hardware device, they have an operating system that is widely used by
72.
by ko_pivot 2y ago
Yes. Character.ai usage is way down from the peak and it is only getting easier for startups to compete in the space. However, the plausibility of an actual FAANG acquisition is almost zero, so this is the next best option… at least for cer
73.
by ko_pivot 2y ago
Yeah the insistence that Windows is insecure by default on purpose is a bit silly, but overall I thought this post was informative.
74.
by ko_pivot 2y ago
The integration you are doing between the LLM and a workflow engine is cool. Are you using something like Temporal or just a database plus a cron job?
75.
by ko_pivot 2y ago
The latest Acquired episode makes a pretty strong case for Ballmer, implying that he led the Microsoft move to own all enterprise software, not Gates.
76.
by ko_pivot 2y ago
Sincere question: how do you actually make this mistake while having the skills to build a web app of this complexity level? All the frontend and full stack frameworks that I’m familiar with try pretty hard to stop you.
77.
by ko_pivot 2y ago
> computers are a very important form of technology. Without it no websites, no Roblox studio, and no nothin. Without computers used have to use tablets. Tablets aren't very useful. Well to do stuff Computers can anyway. Amen.
78.
OpenAI is releasing GPT-4o Mini, a cheaper, smarter model(theverge.com)
6 points by ko_pivot 2y ago|0 comments
79.
Docker without Docker on Fly.io (2021)(fly.io)
1 points by ko_pivot 2y ago|0 comments
80.
Tabloid: A clickbait headline programming language (2021)(tabloid.vercel.app)
217 points by ko_pivot 2y ago|29 comments
81.
Making WarpStream Multi-Region(warpstream.com)
1 points by ko_pivot 2y ago|0 comments
82.
by ko_pivot 2y ago
I think the main issue here is brand and growth. AWS needs to convince CIOs/CEOs to use them over Azure/GCP, not engineers. And even if AWS cared about convincing engineers, we already prefer AWS so introducing new services for c
83.
Nearly all AT&T cell customers' call and text records exposed(cnn.com)
80 points by ko_pivot 2y ago|3 comments
84.
Building and scaling Notion's data lake(notion.so)
1 points by ko_pivot 2y ago|0 comments
85.
The cross AZ cost: how we reduced AWS Data Transfer cost by 80%(medium.com)
2 points by ko_pivot 2y ago|0 comments
86.
by ko_pivot 2y ago
Truly confusing that they would use Firebase for this. Not only is it proprietary as in closed source, but it literally only runs on Google Cloud — can’t even do solid local emulation.
87.
by ko_pivot 2y ago
That’s the thing. I’m not sure they can do much with diffusion models. They need an output that maps to the Figma definition language or whatever it is they call it. In theory an LLM could do this, but they would need to train it on a huge
88.
by ko_pivot 2y ago
Being so visual and using a binary file format, Figma is a great test case in the limits of current gen LLMs. I imagine that in the medium term they will move towards relatively novel approaches rather than the finer tuned LLMs that they ar
89.
by ko_pivot 2y ago
These days AWS primarily does operate their own DCs. This is interesting historically but doesn’t relate to their current footprint.
90.
by ko_pivot 2y ago
I wouldn’t recommend Azure App Service. Container Apps / AKS / Container Instances / Static Web Apps are all fine services for different use cases, but with App Service you are buying into a pretty strange infra approach comp

More ›