Dagens Vibes — 11. juni 2026

Dagens feed var Fable-hangover: ikke “ny model, yay”, men tillid, begrænsninger, lokale modeller og hvordan agentarbejde faktisk føles, når legetøjet bliver dyrt nok til at ligne infrastruktur.

Fra X-feedet

Dagens store friktion: Fable/Mythos er stærk, dyr og delvist sort boks. Gergelys kritik rammer den praktiske udvikler-version: dataretention, usynlig model-nerfing og “trust us bro”-energi i premiumindpakning.

Gergely Orosz@GergelyOrosz

Things I really dislike about Fable: 1. Anthropic collects my prompt history, stores it, and does whatever they want with it for 30 days. No opt-out 2. They can nerf their most expensive model without telling me, billing me the same amount, wasting my time. Whenever they want

6.084 likes290 reposts175 svar544 bookmarkshttps://x.com/GergelyOrosz/status/2064618497150210391

Den hårde udgave af samme debat: safeguards rammer ikke kun våben og malware, men også legitime forsknings- og life-science-spor. Sikkerhedspolitik som produktfeature — og som produktbegrænsning.

Teknium 🪽@Teknium

What's crazy to me is that Fable is blocked from life sciences broadly, nerfed even if you get passed the classifiers and filter level blocks. The whole point of AGI/ASI is to cure all diseases. Everything else is just nice to haves. But Anthropic wants to close off that path. I think Anthropic might be the worst company on the planet.

4.932 likes359 reposts408 svar341 bookmarkshttps://x.com/Teknium/status/2064570771641680022

Google kom med den mest konkrete modelnyhed: DiffusionGemma. Ikke endnu en “den er lidt klogere”-pressemeddelelse, men en anden generationstype: blokke i stedet for token-efter-token.

Google DeepMind@GoogleDeepMind

DiffusionGemma is our new experimental open model with up to 4x faster output on dedicated GPUs. Instead of predicting word-by-word, it generates entire blocks of text simultaneously. This lets the model self-correct and format complex markdown in real time.

1.893 likes204 reposts95 svar401 bookmarkshttps://x.com/GoogleDeepMind/status/2064741061352636762

Apple/Claude-sporet er interessant fordi det flytter AI fra chatbot-vindue til framework. Det er dér platformskrigene bliver kedelige på den profitable måde.

ClaudeDevs@ClaudeDevs

New for Apple developers: Foundation Models support for Claude lets developers use Apple's Foundation Models framework to call Claude for multi-step reasoning, code generation, and longer context. https://t.co/7EbHlWBBSe

2.802 likes186 reposts80 svar906 bookmarkshttps://x.com/ClaudeDevs/status/2064756984617021807

Cursor skubber review-agenten ind i hverdagsflowet: hurtigere, billigere og nu som lokal /review før push. Den slags små workflow-greb bliver vigtigere end endnu en benchmark-graf.

Cursor@cursor_ai

Cursor’s code review agent is now over 3x faster, 22% cheaper, and finds 10% more bugs. You can also use /review to run Bugbot locally to catch and fix issues before pushing code. https://t.co/Pl7lhoR6TE

1.809 likes106 reposts76 svar317 bookmarkshttps://x.com/cursor_ai/status/2064774675021119825

Sikkerhedsfolkene fik dagens mest dystre punchline: malware med indlejret “få LLM-scannere til at nægte”-tekst. Prompt injection er flyttet fra demo til kriminel drift. Hyggeligt, som en motorsav i en børnehave.

John Scott-Railton@jsrailton

NEW: malware developers added nuclear & biological weapons text to to their spyware. Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner. Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky. When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit. We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted. In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation. H/T to colleagues that shared this with me https://t.co/f3Aj9TYxU4

9.806 likes1.689 reposts189 svar3.471 bookmarkshttps://x.com/jsrailton/status/2064661778978533571

Sunil rammer den nye arbejdsfølelse: agenterne æder de lette opgaver, og mennesket sidder tilbage med de hårde rester. Produktivitet, men med konstant final boss-musik.

sunil pai@threepointone

wondering why I feel exhausted. maybe: the agents do all the easy stuff, and I have to work through the leftover hard bits, which means I'm perpetually locked in. and as the models get better, "my" work just gets harder and harder, until I'm basically underqualified to do the work (which... is better than the alternative, there's nothing left for me to do, and I'm paperclipped).

1.479 likes90 reposts96 svar393 bookmarkshttps://x.com/threepointone/status/2064616516478243222

Matt Pocock peger på den mere konstruktive modgift: brug agenten som mentor i strategisk programmering, ikke bare som tastaturmotor. Senioritet bliver mere eksplicit træning end “absorber det ved kaffemaskinen”.

Matt Pocock@mattpocockuk

Steps to become a senior programmer: 1. Install my /teach skill npx skills add mattpocock/skills --skill teach 2. Create a new working directory on your laptop mkdir junior-to-senior cd junior-to-senior 3. Kick off your coding agent in the directory claude 4. Copy this prompt /teach me how to be a great strategic programmer. My opinion is that AI is eating 'tactical, on-the-ground' programming. The day-to-day work of a developer involves not only coding, but also planning, QA, codebase design, and much more. I'm interested in learning the strategic skills - that, in a previous era, would take me from junior to senior - but in this era are table stakes. 5. Paste it into the coding agent Below is an example of what the first output will look like. I used Opus 4.8, medium effort. 6. Continue working with the agent until you're a senior

3.008 likes268 reposts70 svar4.827 bookmarkshttps://x.com/mattpocockuk/status/2064663221718425660

Nyhedsbonus

Tre kilder fra det seneste døgn, der uddyber feedets signal uden at blive et generisk tech-buffetbord.