Øystein Blixhavn's Avatar

Øystein Blixhavn

@blixhavn.bsky.social

Platform advocate @ Omegapoint Norge

134 Followers  |  67 Following  |  52 Posts  |  Joined: 12.08.2023
Posts Following

Posts by Øystein Blixhavn (@blixhavn.bsky.social)

Tja, meme girls er vel ca like riktig

05.01.2026 20:58 — 👍 0    🔁 0    💬 0    📌 0

Any evidence strategy giving inordinate weight to simplistic contextless measures of "developer throughput" in the AI era is fundamentally unserious

17.12.2025 17:01 — 👍 30    🔁 5    💬 3    📌 0

Humble brag. I can only do a pistol squat in my imagination

17.12.2025 16:54 — 👍 0    🔁 0    💬 1    📌 0

Lifehack

18.11.2025 07:15 — 👍 1    🔁 0    💬 0    📌 0
Headline 'Scientists have discovered "third state" between life and death'.

Headline 'Scientists have discovered "third state" between life and death'.

Is it 'online training course'?

17.06.2025 11:47 — 👍 521    🔁 113    💬 12    📌 26
Preview
Cloud Native Bergen Cloud Native Bergen host of Cloud Native Day Bergen and other events. Anything Kubernetes related goes!

Videos from Cloud Native Day Bergen 2025 are now published www.youtube.com/@CloudNative...

31.10.2025 07:51 — 👍 13    🔁 4    💬 0    📌 0

Didn’t know there was any claims of Proton being a honey pot. Not member of either, no

22.10.2025 13:31 — 👍 0    🔁 0    💬 0    📌 0

A what now? No sarcasm

22.10.2025 05:50 — 👍 0    🔁 0    💬 1    📌 0

Bad experiences with Proton?

21.10.2025 19:07 — 👍 1    🔁 0    💬 1    📌 0
Screen cap from linked article, with heading Significance and then text:

Large Language Models (LLMs) are used in evaluative tasks across domains. Yet, what appears as alignment with human or expert judgments may conceal a deeper shift in how “judgment” itself is operationalized. Using news outlets as a controlled benchmark, we compare six LLMs to expert ratings and human evaluations under an identical, structured framework. While models often match expert outputs, our results suggest that they may rely on lexical associations and statistical priors rather than contextual reasoning or normative criteria. We term this divergence epistemia: the illusion of knowledge emerging when surface plausibility replaces verification. Our findings suggest not only performance asymmetries but also a shift in the heuristics underlying evaluative processes, raising fundamental questions about delegating judgment to LLMs.

Sentence starting with "While models often" is highlighted in blue.

Screen cap from linked article, with heading Significance and then text: Large Language Models (LLMs) are used in evaluative tasks across domains. Yet, what appears as alignment with human or expert judgments may conceal a deeper shift in how “judgment” itself is operationalized. Using news outlets as a controlled benchmark, we compare six LLMs to expert ratings and human evaluations under an identical, structured framework. While models often match expert outputs, our results suggest that they may rely on lexical associations and statistical priors rather than contextual reasoning or normative criteria. We term this divergence epistemia: the illusion of knowledge emerging when surface plausibility replaces verification. Our findings suggest not only performance asymmetries but also a shift in the heuristics underlying evaluative processes, raising fundamental questions about delegating judgment to LLMs. Sentence starting with "While models often" is highlighted in blue.

I'd love to see someone try to estimate just how much time and money has gone into research that is either fully undermined by reliance on LLMs or fully pointless --- because obvious if you start from an understanding of what LLMs actually are.

www.pnas.org/doi/10.1073/...

18.10.2025 10:11 — 👍 480    🔁 126    💬 15    📌 14

Har du prøvd å legge deg tidligere?

15.10.2025 07:34 — 👍 0    🔁 0    💬 1    📌 0
Post image Post image

Robin Williams' daughter has some quality thoughts on AI slop

06.10.2025 20:48 — 👍 19759    🔁 7432    💬 137    📌 370

Stop being so cool then, jeez

07.10.2025 08:11 — 👍 0    🔁 0    💬 0    📌 0

if you think chatbots are alive or self aware then i think you would be easily fooled by a tunnel painted on a brick wall

05.10.2025 00:41 — 👍 571    🔁 145    💬 10    📌 2

Nice!

24.09.2025 13:53 — 👍 0    🔁 0    💬 0    📌 0
Post image

We are ready for Cloud Native Day Bergen, are you?! 🚀 2025.cloudnativebergen.dev/program

24.09.2025 07:18 — 👍 0    🔁 4    💬 0    📌 0

What’s the right platform?

24.09.2025 13:29 — 👍 0    🔁 0    💬 1    📌 0

Something I've found really fruitful for me as an architectural principle is understanding that you can simplify a complication, but you have to adapt to complexity.

Which means that complicated things can be abstracted over, but complexity requires re-framing and demands novelty.

22.09.2025 17:20 — 👍 25    🔁 3    💬 1    📌 2

As mentioned, both Flux and ArgoCD have webhooks that can be called for immediate reconciliation:)

21.07.2025 09:35 — 👍 0    🔁 0    💬 0    📌 0

Whenever developers complain that meetings take too much of their time I automatically assume they value the experience of being in The Flow over actually doing something productive. But they think being in The Flow is productive all on its own.

12.07.2025 22:17 — 👍 0    🔁 0    💬 0    📌 0

I’m so appointed by this linguistic reasoning!

27.06.2025 06:32 — 👍 2    🔁 0    💬 0    📌 0

Yaml is the PHP of markup languages

12.06.2025 11:59 — 👍 0    🔁 0    💬 0    📌 0
@clairewillett

honestly the best marketing scheme in history is men successfully getting away with calling women the "more emotional" gender for like, EONS, because they've successfully rebranded anger as Not An Emotion

6:00 PM • Aug 12, 2020

@clairewillett honestly the best marketing scheme in history is men successfully getting away with calling women the "more emotional" gender for like, EONS, because they've successfully rebranded anger as Not An Emotion 6:00 PM • Aug 12, 2020

once again I must tap my own sign

05.06.2025 21:13 — 👍 26433    🔁 5994    💬 229    📌 127
A black and white line drawing with the DOOM logo at the top, "CANDLE DEMON" just below, and then an illustration of a smiling demon with horns kneeling to light some candles on top of a crate.

A black and white line drawing with the DOOM logo at the top, "CANDLE DEMON" just below, and then an illustration of a smiling demon with horns kneeling to light some candles on top of a crate.

Thinking about the fan theory that the DOOM universe has a class of demon whose sole purpose is to go around lightning all the candles in hell. I love him

25.05.2025 17:24 — 👍 76    🔁 10    💬 4    📌 0

Honestly looking forward to speaking at NDC today. Did not think I would manage to feel like that so close to showtime 🤡

23.05.2025 07:20 — 👍 1    🔁 0    💬 1    📌 0
Preview
FN-topp: 14.000 spedbarn risikerer å dø på Gazastripen de neste 48 timene 14.000 spedbarn vil dø på Gazastripen om de ikke får nødhjelp de neste 48 timene, sier FNs nødhjelpssjef Tom Fletcher til BBC tirsdag.

The Norwegian Labour Party can’t blame anyone for their current inaction with regards to the 14.000 starving infants in Gaza.

They are running this country all by themselves. What they don’t do in this moment is 100% on them.

20.05.2025 08:44 — 👍 24    🔁 9    💬 1    📌 0

Rrrrrs

19.05.2025 20:49 — 👍 1    🔁 0    💬 1    📌 0

Q&A from a talk I gave last week.

Q: "What do you think is the biggest threat in cybersecurity right now? Is it post-quantum computing? Is it AI?"

A: Fascism. It's fascism.

14.05.2025 04:07 — 👍 2574    🔁 606    💬 17    📌 21

Ooh, can’t wait! Which podcast? And when is it released?

13.05.2025 19:26 — 👍 0    🔁 0    💬 1    📌 0