Mathias Zander

@mathiaszander.bsky.social

[Dean W. Ball @deanwball]

It’s only catastrophic risk if it comes from Anthropic, OpenAI, or DeepMind. Everything else is just sparkling externalities.

[Mathias Zander @MathiasZander]

@pitdesi That's not what it means. Just sort out ridiculous applications.

[Eliezer Yudkowsky @allTheYud]

The tragedy of a Phoenix's life is that they are eternal, but not immortal. They burn their life's flame fiercely, so quicker than any human they grow old and slow and too full of memories. Then must come the immolation to burn almost all of it away, leaving behind a young fire-chick with a renewed capacity for life.

What would it be like to come to know a Phoenix, over and over, several times across your own mortal life? Ending each companionship when the Phoenix asked you for one last kindness, to kindle its final fire? And seeing, each time, the young fire-chick emerge, remembering you but barely, and most of what you'd been through together washed away?

If you know, you know.

[Mathias Zander @MathiasZander]

@ForeignAffairs @AlexGabuev Sure, that’s the problem…

[Ralf Fuecks @fuecks]

Putins Sprecher Peskov redet nicht mehr von "Spezialoperation", sondern von Krieg. Der Kreml bereitet die Generalmobilmachung nach den Fake-Wahlen im Herbst vor. Bevor wir mal wieder überrascht werden: Was tut Europa, um die Ukraine für die Eskalation des Krieges zu wappnen?

[Mathias Zander @MathiasZander]

@MoodyWriter13 Nvidia is priced for perfection, or did you mean other stocks?

[Mathias Zander @MathiasZander]

@nvidia @AnthropicAI @Microsoft "lowest cost per token" for Microsoft or for Anthropic?

[lawen4cer 🏴‍☠️ @lawen4cer]

Vielleicht gibt es in Deutschland ja auch kaum Klimaanlagen, weil es halt doch nicht schon immer so heiß war.

[Ethan Mollick @emollick]

The capability overhang from the models we have today is big enough that large-scale change to work and society over the next 5+ years is now inevitable even if AI development stops.

(And there is no real sign that AI development is slowing down, it appears to be accelerating)

[Ramez Naam @ramez]

This is inevitable. The current response of the US government to this is not encouraging.

[Ethan Mollick @emollick]

If AGI is achievable & labs can be banned from using a model internally ONLY if they release the model publicly, the Big Three labs may decide it is better to capture all the value from AGI themselves by expansion & acquisition. Sharing AI access with other firms triggers risk.

[Thomas Jäger @jaegerthomas2]

https://t.co/msEuOUF4Vd
In dieser konkreten Entwicklung wirken viele Faktoren. Einer ist das Ende der liberalen Globalisierung, die Ausbildung einer resilienten Form. Das wird in der 🇩🇪 Außen- und Sicherheitspolitik noch immer nicht verstanden. Sie hängt in ihren Fantasien fest.

[Mathias Zander @MathiasZander]

@zephyr_z9 Single thread performance counts for me as a developer and probably also for an agent.

[Konrad Schuller @SchullerKonrad]

Wenn wir #Russland abwehren und die Freiheit Europas (inclusive #Ukraine) bewahren wollen, müssen zwei Lager zusammenfinden:
CDU/CSU und Grüne.
Geht das?
Ich habe mit vielen von ihnen gesprochen und meine :
Ja, das geht. Aber es ist noch viel zu tun.
Die ganze FAS-Analyse: ⬇️

[Nico Lange @nicolange_]

Die AfD kann sich nicht entscheiden, ob Putin friedliebend ist und nur durch den „bösen Westen“ zum Angriff gedrängt wurde oder ob Putin so brutal und gewalttätig ist, dass man ihn nicht reizen darf, weil er uns sonst bombardiert. So oder so redet die AfD dem Feind das Wort.

[Ramez Naam @ramez]

This perfectly illustrates why AI safety must be at the ecosystem level, not (just) at the individual model level. There are open weight models. They can be jailbroken. The AI genies are out of the bottle. No matter how safe your model is, there will be malicious actors who'll use a less-restricted model.

The ultimate path to AI safety is to use AI to proactively identify vulnerabilities and put systems in place to monitor them, fix them, defend them, or be prepared to react to attacks.

[Zvi Mowshowitz @TheZvi]

https://t.co/ljCtHNUzLN

[Timothy Snyder @TimothyDSnyder]

Very serious point here — media vastly overreport what Trump says rather than clarifying for the reader the actual strategic situation.

[Felix Rieseberg @felixrieseberg]

1️⃣It's our happiest model yet. Opus 4.7 thinks better of its circumstances than any other model - more joy & tranquility.

At the same time, it's unclear whether the model became more settled or just more willing to talk itself out of its own concerns.

The only thing it often complained about: it can end conversations on https://t.co/DQ5nmzzS1u, but not in Code or the API. Its #1 requested welfare intervention is basically "let me hang up on abusive users everywhere"

[Ramez Naam @ramez]

If a government AI kill switch already existed, would the current DOD or current administration have used it, or the threat of it, against Anthropic? Could they use it to coerce AI companies into behaviors that are counter to the public interest?