Hacker News: OliverGuy

New comment by OliverGuy in "Google has acquired the data of failed US airline Spirit"

OliverGuy — Tue, 18 Aug 2026 19:20:47 +0000

Not like that doesn't already exist, try scanning and printing a bank note (US, not sure if other countries have similar)

New comment by OliverGuy in "Degraded performance for multiple models"

OliverGuy — Tue, 18 Aug 2026 17:15:40 +0000

Can you give some specific technical examples where 4.7+ are making the wrong architectural decisions?

New comment by OliverGuy in "Are AI labs pelicanmaxxing?"

OliverGuy — Thu, 23 Jul 2026 10:11:22 +0000

If you are a bike nerd and like taking photos of your bike(s) you'll almost always try and have it drive train towards the camera, which would be front on the right side. Possibly this is where is this coming from? Generally also true for retail listings too

New comment by OliverGuy in "OpenAI and Hugging Face address security incident during model evaluation"

OliverGuy — Wed, 22 Jul 2026 05:57:19 +0000

Got a source for that?

New comment by OliverGuy in "Texas spent $4.5M in Spy SUVs that scan phones in 3 minutes"

OliverGuy — Sat, 18 Jul 2026 07:36:15 +0000

It's location tracking more than anything, so yes they would most likely be as the use all the same networks as a normal device

New comment by OliverGuy in "GLM 5.2 beats Claude in our benchmarks"

OliverGuy — Mon, 29 Jun 2026 09:43:21 +0000

Probably a bit niche at the moment really. The only people interested in that are us nerds, and the product segment is very adhoc - especially for the local crowd where an epyc, with a bunch of pcie riders and some 3090s on a steel frame is considered optimal

New comment by OliverGuy in "Identity verification on Claude"

OliverGuy — Mon, 22 Jun 2026 11:50:58 +0000

Curl is one of the most audited code bases, so that's not surprising.

Put it in front of a common-ish Python or NPM package and see what it finds, likely going to be a lot more.

New comment by OliverGuy in "Renting a sewing machine from the library"

OliverGuy — Sun, 21 Jun 2026 08:55:12 +0000

To be fair SRAM are particularly a pain to bleed. Shimano, Magura etc are much easier

New comment by OliverGuy in "I wrote 5000 lines of assembly because I was angry"

OliverGuy — Mon, 15 Jun 2026 14:29:11 +0000

It's comical too, jumpy and takes way to long.

Sure it's kinda cool as a gimmick, but it should be like 600-800ms at most if you do want one

New comment by OliverGuy in "DaVinci Resolve 21"

OliverGuy — Thu, 04 Jun 2026 09:28:38 +0000

Ampere make some big 32-128 core server ARM chips that are somewhere around Zen4c ish in performance, although they aren't commonly used outside of the server context. But they do at least exist.

New comment by OliverGuy in "Ferrari Luce"

OliverGuy — Tue, 26 May 2026 07:52:35 +0000

Ferrari have had indicator buttons in all their cars since about 2010

New comment by OliverGuy in "UK government replaces Palantir software with internally-built refugee system"

OliverGuy — Sat, 16 May 2026 07:08:25 +0000

Have you got a source for the s3/glue/redshift bit?

New comment by OliverGuy in "The Greatest Shot in Television: James Burke Had One Chance to Nail This Scene (2024)"

OliverGuy — Mon, 11 May 2026 06:30:35 +0000

AFAIK youtube will stretch the player window to match the aspect ratio of the source media, lots of cinematic content that's a wider than normal (21:9 I think?) ratio that youtube adjusts the player window to fit around without black bars.

They won't ever squash or stretch video though, so this means the original uploader stretched the 4:3 content to 16:9 at some point before upload

New comment by OliverGuy in "Chevrolet Performance eCrate package (400v/200hp)"

OliverGuy — Thu, 07 May 2026 07:37:47 +0000

I get this is a pretty nice product that they won't sell in huge volume, but that is really steep, was expecting something <20k

New comment by OliverGuy in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

OliverGuy — Wed, 22 Apr 2026 21:45:24 +0000

I would love to have a shit load of small (27B dense. 35B MoE) agents running locally and looking at and ingesting every bit of data about me, my life and what I get up to see what sort of correlations it finds. Give a coding agent access to a data lake of events and let it build up its own analytics tooling to extract and draw out information from that data, and present it to me as daily/weekly/monthly summaries.

New comment by OliverGuy in "Migrating from DigitalOcean to Hetzner"

OliverGuy — Sat, 18 Apr 2026 15:04:30 +0000

What's the HA plan?

Sounds like from the requirement to live migrate you can't really afford planned downtime, so why are you risking unplanned downtime?

New comment by OliverGuy in "Ask HN: What Are You Working On? (April 2026)"

OliverGuy — Mon, 13 Apr 2026 13:43:16 +0000

Just started this weekend on https://gitlab.com/get-otter/otter-sdk

Its dbt inspired stream ETL tool (or maybe just the TL?), it currently just has a dev mode that does RabbitMQ to local Parque files while I'm getting the core of it to a place I'm happy with.

It runs SQL models against the incoming messages and outputs the results to one or more output tables. Has a local WAL so you can tune it to have sensible sized output files (or not, if you need regular updates but at the expense of query perf.)

Planning on adding Protobuf messages, Kafka as a source and S3 and Iceberg tables as sinks this week.

Lightly inspired by a some projects at work where a lot of time and effort was spent doing this and resulted in something not very reusable without a lot of refactor work. Feel like the stream -> data lake pattern should be something that is just SQL + Config, same way dbt is for transformations within a data warehouse.

No plans on adding any cross message joins or aggregations as that would require cross worker communications and I explicitly want to keep the workers stateless (minus the WAL of course)

Would really appreciate any feedback on the core concept, esp. if this is something you'd actually use in prod (if it were finished!) Not sure if there is something that does this already that I don't know about, or if this genuinely fills some sort of hole in the exisitng tooling

New comment by OliverGuy in "Anthropic expands partnership with Google and Broadcom for next-gen compute"

OliverGuy — Tue, 07 Apr 2026 12:05:32 +0000

And you know Nvidia can't be constent with one format for FLOPs within a single graph, 1,000,000x faster but comparing FP32 to FP8 or NVFP4 and acting like it's the same.

New comment by OliverGuy in "Issue: Claude Code is unusable for complex engineering tasks with Feb updates"

OliverGuy — Tue, 07 Apr 2026 07:34:31 +0000

settings.json -> global config Env vars -> settings different to your global for a specific project Slash commands / chat keywords -> need to change a setting mid chat

New comment by OliverGuy in "Google's 200M-parameter time-series foundation model with 16k context"

OliverGuy — Tue, 31 Mar 2026 07:27:41 +0000

Edit, it looks like the paper does

TPUv5e with 16 tensor cores for 2 days for the 200M param model.

Claude reckons this is 60 hours on a 8xA100 rig, so very accessibile compared to LLMs for smaller labs