AI Safety South Africa
@AI_Safety_SA
Followers
61
Following
17
Media
3
Statuses
26
Joined September 2025
Sora 2 is surprisingly good at answering benchmark questions. We suspect it's working in conjunction with a text model which provides the answer.
Sora 2 can solve questions from LLM benchmarks, despite being a video model. We tested Sora 2 on a small subset of GPQA questions, and it scored 55%, compared to GPT-5’s score of 72%.
0
1
13
It's the final day to apply for the Cooperative AI Research Fellowship! ⏰ Final Application Deadline: Wednesday October 1, 2025 (anywhere on earth)
🌍 Join a cohort of ambitious researchers in Cape Town for a cooperative AI research fellowship Spend 3 months researching the biggest problems in cooperative AI, with world-class mentorship from Google DeepMind, Oxford, and MIT researchers. See comments for details!
0
2
5
Applications for the Cooperative AI Research Fellowship have been extended by 2 days! ⏰ Extended Application Deadline: Wednesday October 1, 2025 (anywhere on earth)
🌍 Join a cohort of ambitious researchers in Cape Town for a cooperative AI research fellowship Spend 3 months researching the biggest problems in cooperative AI, with world-class mentorship from Google DeepMind, Oxford, and MIT researchers. See comments for details!
0
0
3
⏰ Today is the final day to apply for the Cooperative AI Research Fellowship! Applications close at midnight, anywhere on earth. Application link below⬇️
🌍 Join a cohort of ambitious researchers in Cape Town for a cooperative AI research fellowship Spend 3 months researching the biggest problems in cooperative AI, with world-class mentorship from Google DeepMind, Oxford, and MIT researchers. See comments for details!
1
0
1
Watch our Fellowship Info Session: https://t.co/xyoTetpcBI Find out more and apply here:
0
0
1
8. @xuanalogue, Assistant Professor at @NUSingapore with a PhD from the MIT Probabilistic Computing Project and Computational Cognitive Science lab.
1
0
0
7. @lrhammond, Research Director at @coop_ai, DPhil Candidate at @CompSciOxford and @GovAI_ Affiliate.
1
0
1
6. @ZhijingJin incoming Assistant Professor at @UofT and Research Scientist at Max Plank Institute. @CausalNLP, @NLP4SG.
1
0
0
1
0
0
4. @sahar_abdelnabi AI Security Researcher at @Microsoft, Principal Investigator at @MPI_IS and Independent Research Group Leader at @ELLISInst_Tue.
1
0
0
1
0
0
2. @conitzer, Professor of Computer Science at @UniofOxford and @CarnegieMellon (where he directs @FOCAL_lab), and Head of Technical AI Engagement at @EthicsInAI.
1
0
1
1
0
0
Spend 3 months researching in Cape Town with world-class mentorship from @GoogleDeepMind, @UniofOxford, @MIT, @CarnegieMellon, @UofT, @NUSingapore, @collect_intel and more. Our mentors include:
1
0
0
Less than one week left to apply for the Cooperative AI Research Fellowship! Applications close: September 28, 2025 (AoE) Application link below ⬇️
🌍 Join a cohort of ambitious researchers in Cape Town for a cooperative AI research fellowship Spend 3 months researching the biggest problems in cooperative AI, with world-class mentorship from Google DeepMind, Oxford, and MIT researchers. See comments for details!
1
6
8
Info session: https://t.co/4K75ELvVng Learn more at our website:
luma.com
Format An overview of the fellowship, presented by AI Safety South Africa (~15 minutes) Outlining the key research areas, presented by the Cooperative AI…
0
0
0
Applications for the Cooperative AI Research Fellowship close in less than two weeks! 📅Attend our info session this Thursday, September 18th Applications close: September 28, 2025 (anywhere on earth) ⬇️Link for info session below!
1
0
0