Hacker News: ody4242

New comment by ody4242 in "Gemini 4 Argon"

ody4242 — Wed, 30 Sep 2026 21:17:49 +0000

what is so terrible with their harness? I've been using gemini cli, now use agy, Pi agent harness, and agent (cursor), and my only real issue with agy was the permission handling, but other than that, it was ok.

New comment by ody4242 in "Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows"

ody4242 — Mon, 10 Aug 2026 12:33:35 +0000

then you switch provider, it's not a monopoly

New comment by ody4242 in "Alphabet's cash burn raises alarm for Big Tech as AI spending climbs"

ody4242 — Thu, 23 Jul 2026 15:12:16 +0000

The upfront cost of facility is ~30%, network infra 10-15%, land/utilities is small percentage, power could be significant for an AI DC. The servers are ~50-60% only.

New comment by ody4242 in "Alphabet's cash burn raises alarm for Big Tech as AI spending climbs"

ody4242 — Thu, 23 Jul 2026 14:51:51 +0000

They are building new datacenters for the AI demand, so around half of this CAPEX is not for the GPU-s, and those will not be replaced every 3-5 years.

Also, TPUv2 was introduced in 2018, and still not completely retired in all regions, from accounting pov, it has been written down to 0, but they are still working.

New comment by ody4242 in "Alphabet's cash burn raises alarm for Big Tech as AI spending climbs"

ody4242 — Thu, 23 Jul 2026 14:34:33 +0000

This is a good chart that shows historical CAPEX spending. Hyperscalers have been through a couple CAPEX cycles like this, they all know what they are doing.

https://eco3min.fr/en/big-tech-capex-revenue-ratio-quarterly...

New comment by ody4242 in "Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber"

ody4242 — Wed, 22 Jul 2026 14:20:51 +0000

Why would they go into price war when they are compute constrained?

New comment by ody4242 in "Turn your best AI prompts into one-click tools in Chrome"

ody4242 — Tue, 14 Apr 2026 19:55:07 +0000

try to add it here: https://gemini.google.com/saved-info

New comment by ody4242 in "Big Data on the Cheapest MacBook"

ody4242 — Thu, 12 Mar 2026 14:40:50 +0000

Idk how an AWS region would respond to a power outage, but i have tested this in AWS Outpost, and there, if you power down a rack, then power it back again, the baremetal instances will not be recreated. (I was surprised as I was expecting the EC2 health check to terminate them, but it does not work like that.) My understanding is that if you stop/start an instance, your local storage is gone (as the instance might even end up in a different host), but if you just reboot the instance, it should keep the local storage.

New comment by ody4242 in "Big data on the cheapest MacBook"

ody4242 — Thu, 12 Mar 2026 12:49:13 +0000

I would have benchmarked with an instance that has local nvme, like c8gd.4xlarge.

New comment by ody4242 in "Gemini 3 Deep Think"

ody4242 — Fri, 13 Feb 2026 13:30:44 +0000

I'm 100% sure that all providers are playing with the quantization, kv cache and other parameters of the models to be able to serve the demand. One of the biggest advantage of running a local model is that you get predictable behavior.

New comment by ody4242 in "The Waymo World Model"

ody4242 — Sat, 07 Feb 2026 10:56:24 +0000

Yesterday GPT 5.2 wrote a python function for me that had the import in the middle of the code, for no reason. (It was a simple import of requests module in a REST client...) Claude I agree is a lot better for backend,Gemini is very good for frontend

New comment by ody4242 in "Euro firms must ditch Uncle Sam's clouds and go EU-native"

ody4242 — Sat, 31 Jan 2026 14:32:10 +0000

You don't need hundreds of services. Give me virtual machines, reliable block storage, file storage and object storage, networking, dns, managed kubernetes, and it will cover the majority of workloads in Europe that run on Openshift or Openstack today.

New comment by ody4242 in "AWS is 10x slower than a dedicated server for the same price [video]"

ody4242 — Wed, 26 Nov 2025 09:02:10 +0000

Sure, EBS or any network-attached storage is expected to be a lot slower than a local SSD for synchronous writes or random reads, as there is a network stack in between. But my understanding is that for those usecases, you can use metal instances with local nvme. (ephemeral though)

New comment by ody4242 in "The DOJ still wants Google to sell off Chrome"

ody4242 — Sat, 08 Mar 2025 22:23:26 +0000

MS does sell ads, they have an advertising platform.

New comment by ody4242 in "AWS Makes Cloud Formation Stack Creation Up to 40% Faster"

ody4242 — Mon, 18 Mar 2024 09:05:00 +0000

Terraform

Crossplane (although it's quite new, and missing a lot of things that Terraform has)

New comment by ody4242 in "Auto industry executives admit electric vehicle plans are in jeopardy"

ody4242 — Fri, 03 Nov 2023 16:42:33 +0000

The Model3 has a dedicated button for the windscreen wiper on the steering wheel, and then you can adjust with the scroll buttons. It certainly has a learning curve, and it would be better if you could rely on muscle memory for these things.

I usually spend a couple of minutes when I drive someone else's car to get familiar with the controls, as the rear window windscreen wiper, light controls (especially fog light) can be tricky on old cars as well.

New comment by ody4242 in "Ubicloud – open, free and portable cloud"

ody4242 — Thu, 17 Aug 2023 07:15:48 +0000

What makes you stand out from this list of other managed openstack providers? https://www.openstack.org/marketplace/public-clouds/

New comment by ody4242 in "Real-Time Air Quality Index Visual Map"

ody4242 — Wed, 05 Jul 2023 18:52:40 +0000

I have seen similar numbers from multiple sensors across our city during the fireworks at New Years eve.

New comment by ody4242 in "Tesla FSD data is getting worse, according to beta tester self-reports"

ody4242 — Wed, 14 Dec 2022 18:56:07 +0000

Imho, if Tesla could train a huge neural network to demonstrate that self driving is possible with their sw stack, they would have done that already to let Elon Musk have it's road show with it's capabilities.

(Running ChatGPT does not require a supercluster, training the model does)

New comment by ody4242 in "Reddit's favorite products in one place"

ody4242 — Thu, 08 Sep 2022 15:49:10 +0000

Where is wallstreetbets? :)