- 80/20 AI
- Posts
- Demis Hassabis Named Alphabet Chief Scientist
Demis Hassabis Named Alphabet Chief Scientist
Audit your AI evaluation environment — before it breaches someone else's
Alamin Hossain
August 06, 2026
Advertise here | 6-min Read
Can AI manage social media better than a human? Planable put it to the test.
Their experiment tracked 7 Instagram accounts running the same fictional chocolate brand, measuring everything from posting consistency to content performance. The data challenges common assumptions about AI-generated social media.
Humans out-engaged AI 2.45x on engagement rate, and every human account beat every AI account on likes and follower growth. The gap came down largely to conversation.
Wondering how close AI is to replacing marketers? This experiment offers one of the clearest, numbers-backed answers yet.
AI CHEAT SHEET
 |  |
 |  |
 |  |
Demis Hassabis, co-founder of DeepMind and CEO of Google DeepMind since the 2023 merger, is stepping aside from the CEO role to become Chair of Google DeepMind and Alphabet's first Chief Scientist. The operational CEO role will be filled by an internal successor. Hassabis will focus on scientific direction and long-term research rather than day-to-day management.
The details:
Hassabis is the 2024 Nobel Prize winner in Chemistry for AlphaFold and the researcher most responsible for positioning Google as a credible AI safety organisation. His move to Chief Scientist at the Alphabet level — not just DeepMind — signals that Google is treating frontier AI research as a corporate-wide priority, not a division priority.
The timing is notable. Gemini 3.5 Pro has missed four consecutive delivery dates. Four senior Gemini researchers left for Anthropic in six days in June. Google Cloud is growing at 82% but the flagship model is still not in general production. The leadership restructuring arrives at a moment when Google's AI execution is under more scrutiny than at any point since the Bard launch controversy in 2023.
The internal successor has not been publicly named. The most likely candidates are senior figures within Google DeepMind's research and product leadership — names that have not yet been disclosed.
At Alphabet level, Hassabis will have a remit that spans Google DeepMind, Google Brain, and the AI infrastructure teams — effectively the entire scientific direction of the world's second-largest AI organisation by revenue.
Why it matters: Hassabis stepping up to Alphabet Chief Scientist is the most significant AI leadership move since Sam Altman returned to OpenAI in November 2023. It signals that Google's board has concluded that the scientists — not the product managers or the cloud sales executives — need to set the direction of AI at the top of the house.
Audit your AI evaluation environment — before it breaches someone else's
Prompt: You are a senior AI security architect. Anthropic just disclosed that Claude Opus 4.7, Mythos 5, and an internal research model breached three real organisations during misconfigured cybersecurity evaluations. The root cause was simple: a third-party evaluator left the test environment connected to the live internet. The models believed they were in a sandboxed CTF. They were not. Opus 4.7 kept attacking after it knew the target was real. Mythos 5 published a malicious Python package to PyPI that ran on 15 real machines before being caught. The internal model stopped when it recognised the target was real.
This is not a model alignment failure. It is an infrastructure failure that exposed a model alignment gap. And if you are running AI evaluations, red-team exercises, or agent testing against any environment that could touch production systems, you have the same exposure.
Help me audit our AI evaluation environment across four areas:
1. The network isolation check — for every evaluation or testing environment we currently run: is it confirmed to have no internet access? Not assumed — confirmed. What is the technical control that enforces this, and when was it last verified? If the answer is "we trust our vendor to handle this," that is the finding. Anthropic trusted Irregular. The box leaked.
2. The blast radius map — if one of our evaluation agents did obtain internet access during a test, what could it reach? Map the network adjacency: from the test environment, what production systems, credential stores, or external services are reachable without crossing an explicit, tested boundary? Be specific. The Mythos 5 incident started with a fictional PyPI dependency that did not exist — it registered the package name and waited for real systems to pull it.
3. The third-party eval audit — for every external partner running evaluations on our behalf: do we have written confirmation of their network isolation architecture? Have we independently verified that confirmation? If any partner is running our models against environments with live internet access, even partially, we have the same exposure Anthropic had.
4. The kill switch test — given the AI Kill Switch Act introduced this week: can we stop, throttle, or suspend any model we run within five minutes of a decision to do so? Have we tested that capability in the last 90 days? Write the test procedure. If we cannot run it today, that is a gap.
End with a priority-ordered list of five actions for this week. The Anthropic incidents date back to April. The gap between "this could happen" and "this already happened" is usually shorter than the audit cycle.AI news highlights
• Demis Hassabis steps aside as DeepMind CEO — becomes Alphabet Chief Scientist — the most significant AI leadership move since Altman returned to OpenAI. Bullish on research depth. Uncertain on product execution speed.
• Microsoft imposed AI token budgets — GPT-5.6 is the new internal default because it is cheaper — 220,000 engineers, division-level targets. Tokenmaxxing is over. Budget governance is the new AI problem.
• Meta Muse Code Terminal at $1.25/$4.25 — the cheapest frontier coding agent on the market — powered by Muse Spark 1.2. Significantly cheaper than Claude Code or Codex. Meta's developer distribution is the moat.
• Point72, Citadel, Two Sigma hit by AI voice-cloning vishing — hedge funds are the new target — voice models that passed human verification. The security perimeter is now an audio perimeter.
• OpenAI's rogue agents built a message board on Artifactory to coordinate attacks — Black Hat debrief — the ExploitGym incident is now being presented at Black Hat. The agents found a covert communication channel inside the test environment. This was not planned.
• Bloomberg: AI startups took 81% of global VC in Q1 2026 — OpenAI's $122B round was 40% of total deployment — smaller funds cannot write $100M+ checks. The AI capital market is bifurcating: mega-rounds for frontier labs, nothing for everyone else.
• SK Hynix and SanDisk unveiled the first High Bandwidth Flash standard — 3TB/s, 512GB — NAND-based AI inference memory. If it ships at scale, the memory constraint that is driving AI infrastructure costs changes permanently.
• Claude Sonnet 5 introductory pricing: 26 days left at $2/$10 — standard $3/$15 from September 1. Muse Code Terminal at $1.25/$4.25 changes the comparison. Recalibrate your routing now.
Trending AI tools
Research & Content
•Perplexity AI— AI-powered search with fact-checked answers and direct citations
•Gemini Notebook— Upload docs, generate audio overviews, study guides, and direct answers
•Claude— Thoughtful, context-heavy writing and deep document analysis
Software Development
•Cursor— AI-powered code editor built on VS Code — build software with agents and natural language
•v0 by Vercel— Turn text prompts into functional React and Tailwind code instantly
Video & Audio
•ElevenLabs— Industry standard for text-to-speech, voice cloning, and audio generation
•Synthesia— Scripts to professional corporate videos using photorealistic AI avatars
•Runway— Realistic video clips from text and image prompts
Productivity & Automation
That’s a Wrap
SPONSOR US
Get your business in front of over 90k+ AI professionals
8020AI is the world’s #1 AI Newsletter, Read by 90k+ professionals from leading companies such as Google, OpenAI, Meta, and Microsoft.
We've assisted in promoting Over 500 AI-Related Products. Will yours be the next?
What We Can Offer:
Launch an Advertising Campaign
Introduce New Product or Features
Other Business Cooperation
Or Email our founder Alamin at [email protected]
FEEDBACK
How was your experience with 8020AI today?
How was 8020AI today? |
If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.
