My best interview in some time.
Rohin Shah leads AGI alignment/safety at DeepMind.
And he has a lot of spicy personal takes:
We probably won’t get catastrophic misalignment (00:49)
Safety 'commitments' have severe limitations (10:38)
The intelligence explosion probably isn't imminent (1:52:44)
Why he's not working to pause AI advances (51:44)
Pre-deployment evals aren't the right focus (for catastrophic risks) (37:41)
Signalling concern for safety sometimes diverts resources from actually making AI safe (01:09:51)
Reading AI thoughts is v useful for safety – and we'll probably be able to for years to come (54:17)
Governance is somewhat more likely to be the bottleneck than alignment (43:55)
Rohin's team doesn't have a veto, and that's OK (27:36)
Central banks are a promising model for regulating AI (33:34)
Also:
Google DeepMind's actual plan for building AGI safely (1:40:29)
How external researchers can positively influence big AI companies (2:21:55)
The roles GDM most needs to hire for (2:37:03)
On the 80,000 Hours Podcast. Links below - enjoy! (@rohinmshah)
@NtanyaPM Did Africans not have proper family structures 250 plus years ago before European surname changes and western customs that validate marriages... where Africans failing to uphold families ?
Where African weddings inauthentic before white wedding customs?
@NtanyaPM I advice both you and your friend to stay away from people's ambitious and hardworking daughters and perhaps marry or date people in your intellectual spec who share the same *values* as you... do not go for girls with a bright future and try to dim their light
@NtanyaPM If the family structure can be devalued by the lack of a European custom of surname change ? Then perhaps that structure was weak to begin with
@NtanyaPM Please respectfully ask your friend to let people's child go ... the signs are already present that he will be a source of grief to her throughout their marriage because of his insecurities that are not even originally of african descent...
AMLD Africa is coming back.
A few days bringing together researchers, builders, students, and industry leaders to shape the future of AI on the continent.
⏳ 19 days to go.
🎟 https://t.co/FggGWLwQlm
Only a few left.
#AMLDAfrica#TechEvents#Johannesburg#WitsUniversity
She started with zero tech skills, today she’s a tech tutor.
With a background in Accounting and Finance, Munenyashaishe Hove wanted more. During the COVID-19 lockdown, she decided she needed a career she could thrive in from anywhere. That decision led her to the Women Techsters Fellowship.
In this episode of #TechEmpowered, Hove shares how the program became her turning point, giving her the skills, confidence, and clarity to transition into tech. Today, she’s a Data Science & AI engineer, instructor, researcher, and author.
Her story proves it’s never too late to rewrite your career and succeed.
🎧 Listen now: https://t.co/maJgOMGqgs
#TechEmpowered #WomenInTech #TechJourney #WomenTechsters #Tech4Dev #CareerTransition #DigitalSkills #DataScience
@GlobeEyeNews Sam Altman “If you go talk to ChatGPT about your most sensitive stuff and then there's a lawsuit, we could be required to produce that … It makes sense to … really want the privacy clarity before you use it a lot.”
You may have heard about AI 2027—the scenario that got attention from US VP JD Vance.
But you haven't seen it like this: a breakdown of the scenario, its implications, and how a sane world would respond.
Do we need a new AI Bill of Human Rights? Join leading experts in a discussion, convened by the @EthicsinAI_AFP of the @EthicsInAI with the Center for Transnational Legal Studies, led by Professor Yuval Shany Including Professor Alondra Nelson, in London 📷 10 July 2025, 17:30
📢Publish your work in the ACM Journal on Responsible Computing, for a special issue focused on #TrustworthyAI and Autonomous Systems. @TheOfficialACM@ACMDL@gramchurn
Details: https://t.co/YnDOSCjIot
📅 Submission deadline: 10 Oct 2025
👉 For more: https://t.co/cfN9hdn7VV
#LetFreedomSing A Holistic Approach to Social Justice, by Brian Tamuka Kagoro
Fighting for social justice is not inconsistent with seeking transformative, sustainable , and inclusive development.
Valuing science, data, and evidence should never make one blind to the racialized socio-economic, political, and cultural contexts that determine prioritisation of certain types of data or give it particular nuances . Data is not neutral, nor is it always a gender , generational, or race-blind product .
Science? Science historically claimed that blacks were neanderthal cousins until recent genetic research contradicted this long-held claim. Racial superiority, like make superiority, and other prejudices were for centuries backed up by supposedly "scientific evidence." Science - like data - in the hands of bigots , zealots , and criminals can easily be weaponised , perverted , commercialised, and instrumentalized to achieve non-scientific ends .
Technology is designed by humans . AI is programmed or fed by human beings with very particular social orientations, biases , beliefs , and ethics . If you do not thoughtfully regulate it , you risk becoming a victim of human excesses and prejudices institutionalised through technology and AI .
We need to retain 360-degree political, social, and economic vision and perspective.
1. Don't be against enterprise, entrepreneurship, business, and innovation, but oppose exploitation. Business is not necessarily synonymous with exploitation
2. Passionately and sustainably fight indigenised oppression and oppressors, Kleptocrats and liars , but also understand the more complex and devastating interests of empire and external exploitation and exploitors . Fight these with three times the energy, passion , and knowledge.
3. Seek transformative development, but ensure that it re-humanizes communities and does not impoverish them culturally and economically
4. Support developing countries and their leadership to gain greater voice and influence in multilateral spaces , but don't condone corruption, impunity , and flagrant violations of International humanitarian and human rights norms that enable people to enjoy their freedoms and access justice
5. In all things, adopt a balanced approach and seek to understand the deep nuances . When you support a position , solution, or person, have an in-depth understanding of the weaknesses, shortcomings, limitations , and potential dangers of the same . Don't become an activist with little nuanced absolutist views . Everything in life requires balance to avoid toxicity and self-destruction.
#apn2024assembly
Day 2 kicks off with a talk on “Decolonizing Knowledge and Centering African Wisdom” by Brian Kagoro. He is deconstructing western appropriation of African knowledge and exposing…