Hacker News: nolist_policy

New comment by nolist_policy in "Btrfs/ZFS/bcachefs under workloads classic benchmarks skip"

nolist_policy — Sun, 20 Sep 2026 07:15:54 +0000

Checking the bus factor:

Btrfs regulars:

- 1 from Meta

- 1 from Oracle

- 4 from SuSe

- 2 from WDC

bcachefs:

- Kent Overstreet

New comment by nolist_policy in "Btrfs/ZFS/bcachefs under workloads classic benchmarks skip"

nolist_policy — Sun, 20 Sep 2026 07:00:21 +0000

Not true, Boris Burkov from Meta works on btrfs. And two people from WDC.

New comment by nolist_policy in "Btrfs/ZFS/bcachefs under workloads classic benchmarks skip"

nolist_policy — Sun, 20 Sep 2026 06:47:50 +0000

Only if you have quotas enabled IIRC.

New comment by nolist_policy in "Qantas Airbus A380 engine failure in 2010 (2023)"

nolist_policy — Thu, 03 Sep 2026 07:03:05 +0000

Not quite: https://en.wikipedia.org/wiki/Food_libel_laws

New comment by nolist_policy in "Creating Backup Storage Sucks"

nolist_policy — Thu, 03 Sep 2026 06:51:35 +0000

> Sync, rotating snapshots, and deduplicated backups need to be kept entirely separate if you want any hope of maintaining your sanity.

Not if you use git-annex.

New comment by nolist_policy in "GLM-5.3 is now open-weight"

nolist_policy — Sat, 29 Aug 2026 07:51:03 +0000

Yes, The two couldn't be more different.

GPT-3 is probably one of the last models trained without synthetic data. Meanwhile gpt-oss is rumored to be only trained on synthetic data.

I also miss the 2020 GPT-3 based AI Dungeon fine tune.

New comment by nolist_policy in "GLM-5.3 is now open-weight"

nolist_policy — Sat, 29 Aug 2026 07:09:02 +0000

Then take DeepSeek V4 flash with AA score 52. Runs unquantized on 2x DGX spark with 1M context.

New comment by nolist_policy in "Omacom Foundation funding hits $10M"

nolist_policy — Mon, 24 Aug 2026 12:15:00 +0000

I don't get why they choose a tiling window manager.

New comment by nolist_policy in "Omacom Foundation funding hits $10M"

nolist_policy — Mon, 24 Aug 2026 12:01:20 +0000

It's what happens when tech bros run a Linux distro https://omarchy.org/news/2026/09/the-omarchy-core-team/

New comment by nolist_policy in "I'm becoming AI-blind"

nolist_policy — Sat, 22 Aug 2026 08:25:36 +0000

I think you can also just use the random number generator (seeded with a secret key) as the watermark. Then the probability distribution is exactly the same.

New comment by nolist_policy in "Omacom Foundation launches with $8M"

nolist_policy — Fri, 21 Aug 2026 17:49:52 +0000

Omarchy is not a product though.

New comment by nolist_policy in "c100"

nolist_policy — Fri, 21 Aug 2026 16:54:35 +0000

Only for non-commercial usage tho.

New comment by nolist_policy in "c100"

nolist_policy — Fri, 21 Aug 2026 16:38:12 +0000

So from these pictures they just put an off the shelf minipc in a aluminium case and ship it with a matching keyboard.

New comment by nolist_policy in "Ox Alpha"

nolist_policy — Fri, 21 Aug 2026 07:27:58 +0000

AI labs are limited by the available training data, the best way to improve model performance is more and better training data.

The AI labs and the downstream companies that sell training data to them vacuum up everything they can.

Illegal residential proxies (botnets) that once have been used by hackers and scammers are now used to vacuum up the Internet.

They are now vacuuming up antique books that are practically useless.[1]

In face of this is is unthinkable to me that they are not training on API data.

> The business loss of trust would outweigh any benefits of the data.

The loss of trust is already here.

I know of one German company that uses AI only in areas where they have to compete with (foreign) startups. For their core business and everything else they are waiting for an on-prem solution. Apparently Microsoft can provide on-prem GPT-5.

[1] https://lesekauz.de/forum/thread/1999-sammelbestellungen-von...

New comment by nolist_policy in "Ox Alpha"

nolist_policy — Fri, 21 Aug 2026 06:46:46 +0000

Everyone trains on your data.

With Chinese providers at least I'm getting a open weight model out of it.

New comment by nolist_policy in "DeepSeek V4 Pro 0813"

nolist_policy — Wed, 12 Aug 2026 17:09:19 +0000

DeepSeek V4 Flash is the "too cheap to meter" of AI. And you can run the full unquantized model locally for $8000 (2x DGX Spark) at full 1M context and decent speeds: https://github.com/elsung/dgx-spark-deepseek-v4-flash#-long-...

New comment by nolist_policy in "AMD acquires Taalas to boost inference performance by etching models in silicon"

nolist_policy — Fri, 07 Aug 2026 07:24:25 +0000

AMD doesn't even have their own fans.

New comment by nolist_policy in "Cloudflare OS: an open platform for agents, apps, and work"

nolist_policy — Thu, 06 Aug 2026 11:47:29 +0000

Let me quote the workers GitHub:

> WARNING: workerd is not a hardened sandbox

> workerd tries to isolate each Worker so that it can only access the resources it is configured to access. However, workerd on its own does not contain suitable defense-in-depth against the possibility of implementation bugs. When using workerd to run possibly-malicious code, you must run it inside an appropriate secure sandbox, such as a virtual machine. The Cloudflare Workers hosting service in particular uses many additional layers of defense-in-depth.

Sandstorm was great because it did proper sandboxing. This is pretty weak by comparison.

New comment by nolist_policy in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"

nolist_policy — Thu, 06 Aug 2026 07:16:11 +0000

> Anyway, the actual theory is that Google and Meta have fallen behind because they've been playing by this playbook of focusing on scale and training data, whereas OpenAI and Anthropic have done so well because they are likely doing much more interesting things to improve their models over time.

Google Gemma disproves this

New comment by nolist_policy in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"

nolist_policy — Thu, 06 Aug 2026 06:48:56 +0000

> Google probably now fumbled with world models [...].

Can you explain this more precisely?