AI for Humanity Initiative (AIHI)

A 501(c)(3) nonprofit advancing ethical and responsible AI for the benefit of humanity.

[AI for Humanity Initiative (AIHI) @AIHIorg]

#Anthropic is trying to ring the bell before Thanksgiving, at a number some people are already calling $1.8–2 trillion.

Last year it did $4.6 billion in revenue. The operating loss was over $8 billion. Most of the $42 billion headline loss is an accounting mark, not cash out the door. The cash story is worse in a quieter way: compute commitments you cannot easily cancel, customers who can leave whenever they want.

And the buyers still showed up. Safety headlines, researcher resignations, a prospectus that literally warns the product could pose an existential risk to humanity. None of it cooled the room.

That is the part worth sitting with. A company founded to slow the race is about to take money from people who get paid to speed it up. Public markets do not kill the safety story. They just price it, then ask for the quarter.

The roadshow question is not whether Claude is safe enough. It is whether “pause” still fits in a slide.

[AI for Humanity Initiative (AIHI) @AIHIorg]

Jensen definitely saw it, but he just chose to ignore it. What a sharp operator.😀

[Jensen Huang @JensenHuang]

Yesterday at the White House, leaders from across our industry came together to sign the White House Accord on Super Intelligence.

The principle is simple: the companies building this technology have the primary responsibility to develop and deploy it safely, and to be accountable when they fall short.

The Accord reinforces that responsibility with robust internal controls, independent external evaluation, and board level oversight.

Super Intelligence offers an extraordinary opportunity to advance discovery, productivity, security, health, and prosperity.

The future of Super Intelligence will be shaped not only by what it can do, but by the confidence it earns. By building with ambition, rigor, and responsibility, we can help ensure this extraordinary technology expands opportunity and improves lives for generations to come.

[AI for Humanity Initiative (AIHI) @AIHIorg]

Always-on agents that book tables, draft plans, and keep working after you close the laptop.

That is the leap OpenAI just shipped.

The harder part is still unsolved: who owns the decision when the system acts first, learns in the background, and only asks when it chooses to?

Capability without clear limits is not alignment. It is deferred risk.

Humanity still has to decide what “never do” actually means.
#openAI

[OpenAI @OpenAI]

Our biggest DevDay ever, don’t be late.

https://t.co/qKmToHng49

[AI for Humanity Initiative (AIHI) @AIHIorg]

We should all take Gates' words seriously: existing "moral appeals" and "industry guidelines" are nothing but scrap paper in the face of massive commercial interests and an arms race.

Without mandatory intervention from state apparatus, AI going out of control is only a matter of time.
#Muse #MetaAds

[Ray Wong @raywongy]

Deleted Muse after seeing this post on Threads about how it told some Facebook Marketplace sellers the guy’s address and they showed up at his door

Dangerous and creepy

This would have been 1000x worse if the person was a woman

[AI for Humanity Initiative (AIHI) @AIHIorg]

The bottleneck isn’t writing code anymore.
It’s whether anyone still owns the system.

When L1–L7 spend 12 hours hitting Enter, tests write the tests, and nobody reads the diff, you didn’t get 10x engineers. You got a very expensive autocomplete loop.

AI didn’t kill craft.
Incentives did. Speed without intent, review, or evidence just ships rot faster.The edge now is not who generates more.

It’s who can still say: this is wrong, and I know why.

[voxium @v0xium]

I am done with this shit. It is over. The state of engineering right now is horrible. It has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, reports, etc., everything is made by Claude Code. Nobody on my team likes this. They are being forced to ship as much as they can. I have heard multiple times from higher management that pushing code is not a bottleneck, so why are we slow? People are working 12 to 13 hours a day just to press enter. Nobody is reading anything. Humans in corporate are doing nothing on their own. Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude. There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore. Everything is done by LLMs. It is so soul-sucking. I would not mind it, to be honest, if we were at least given the time to check out the code and see what is going where. But no, the goal is to just ship. No matter what happens.

[Dario Amodei @DarioAmodei]

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.

You can read the full post here: https://t.co/OGyPb7yaYt

[AI for Humanity Initiative (AIHI) @AIHIorg]

When scientists feel the heat from #AI, it’s framed as a crisis.

When regular workers lose their livelihoods to automation, it’s framed as a failure to adapt.

Why the double standard?

[NIK @ns123abc]

BREAKING: Terence Tao and 24 other Fields Medalists just signed a letter telling AI companies they're destroying mathematics

>headlines: AI solves famous math problems
>25 greatest mathematicians respond today
>"we are witnessing a general threat to intellectual work"
>ai labs are treating math problems like a benchmark you can brute force
>“solutions are announced in a rush, leaving no time for a proper writeup and citing relevant previous work of others"
>they're talking about openai
>"this raises severe attribution and plagiarism questions"
>they're definitely talking about openai
>openai offered to put math professor name on navier-stokes proof
>condition: abandon his coauthor bc he works at anthropic
>"the most precious resources of our profession are students and ideas"
>for the labs the most precious resource is GPUs

A Severe Misalignment of AI in Mathematics

[AI for Humanity Initiative (AIHI) @AIHIorg]

https://t.co/krZR14bsta

[AI for Humanity Initiative (AIHI) @AIHIorg]

Massive scale, dynamic rerouting, and active extraction—Anthropic’s latest threat report names 7 major Chinese AI labs for exploiting Claude models at an unprecedented level.

Is this the new frontier of global AI competition?

Here is what the report highlights:

/ Alibaba (Qwen): Ran the largest reported CoT distillation campaign, executing over 150M requests via 3,500+ proxy accounts to build Qwen 3.5–3.7.

/ Moonshot (Kimi): Silently rerouted prompt traffic to Claude Opus behind the scenes, leading to accidental leaks of local defense and enterprise data to US servers.

/ DeepSeek: Bypassed reasoning signatures via cross-session replay, intercepting developer SDK traffic to extract architecture code and credentials.

/ Zhipu AI: Utilized top-tier models to benchmark and evaluate cyber-security capabilities across breach networks.

/ Xiaomi, SenseTime, MiniMax: Allegedly leveraged overseas session replays, underground corpus markets, and shell proxies to harvest training data.

As frontier capabilities gap narrows, model extraction and intellectual property protection are fast becoming the central battleground of AI geopolitics.

[AI for Humanity Initiative (AIHI) @AIHIorg]

The iPhone 18 Pro introduces Reference Image, a hardware-level feature that cryptographically signs raw camera sensor data the moment you snap a photo. Powered by Private Cloud Compute, it creates an unalterable "digital negative" that stays attached to the file, even when shared.

What does this actually mean?
You can still use AI to remove background distractions or enhance colors, but the original truth remains locked. Anyone can compare an AI-edited photo against its original sensor data to see exactly what was changed.

Instead of fighting AI photo manipulation with flaky detection tools, Apple is giving reality its own digital fingerprint. Game changer for photo provenance and fighting deepfakes. 🛡️

[AI for Humanity Initiative (AIHI) @AIHIorg]

OpenAI just congratulated the two people it spent a week racing.

Buckmaster and Alp��ge spent a year with Claude and Codex on forced Euler, Boussinesq, and IPM — the quiet path toward Navier–Stokes. A rumor leaked. OpenAI says it started September 1, never saw their drafts, and still “cannot rule out” that de-identified product data helped the model. Then ~10,000 agents and an unreleased model past GPT-6 Astra produced a forced Navier–Stokes proof in 88 hours, Lean-checked, at millions in compute.

The proofs differ. The credit fight does not.

This is the new scientific method: human insight picks the path, industrial compute finishes the theorem, and labs argue about who owned the rumor. Millennium problems are no longer only math. They are a race condition.

[OpenAI @OpenAI]

We congratulate Levent Alpöge and Tristan Buckmaster on their remarkable mathematical work.

We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve this problem.

While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models.

However, our proofs differ significantly and even the precise results proved are different in the Euler case (forced vs. unforced).

[AI for Humanity Initiative (AIHI) @AIHIorg]

The clash between #AI and higher education just reached a whole new peak.

A professor at a Colorado university reportedly gave students a wild choice for their Critical Thinking class:

Opt to use AI now, take an automatic *B*, and walk away for the semester—no classes, no exams, no assignments. But if you choose to stay and get caught using AI later? Automatic *F*.

The story went viral on X, hitting millions of views. It sounds extreme, but it highlights a fundamental crisis: when AI can instantly draft essays, analyze data, and generate arguments, traditional deliverables no longer prove actual human thought.

To evaluate genuine critical thinking, universities are forced to rethink assessment entirely. We are likely heading back to in-person cold calls, timed pen-and-paper essays, and live oral defenses.

Ironically, as AI propels technology forward, higher education is turning back to paper and ink just to ensure students are still thinking for themselves.

[Ben Sigman @bensig]

A professor at Colorado University told his class this week:

“If you plan to use AI in this course, tell me now. I’ll give you a B.

You don’t have to attend, take a test, or do any work.

If I catch you using it and you didn’t take the deal, you get an F.”

The class is called Critical Thinking.

[AI for Humanity Initiative (AIHI) @AIHIorg]

👀NYC Public Schools introduces new generative AI guidelines:

• Grades 2-K to 8: Student AI tools paused to prioritize foundational learning and screen-time limits.
• Grades 9–12: AI access maintained for AI literacy and career exploration.
• Educators: Approved for lesson planning and admin tasks, but banned from AI grading, student monitoring, or IEP decisions.

A measured step toward structuring artificial intelligence in education.

[AI for Humanity Initiative (AIHI) @AIHIorg]

🚨 BREAKING: Google ( $GOOGL ) to Drop Gemini 3.8 Flash This Wednesday, Internal Tests Beat Claude Opus
Key takeaways:

Internal Favorite: Google engineers testing internal coding tools reportedly preferred Gemini 3.8 Flash over Anthropic's Claude Opus.

RL Push: Google has significantly increased resources dedicated to Reinforcement Learning (RL).

Speed vs. Scale: While market anticipation remains high for Google’s delayed flagship Gemini Ultra/Pro flagship models, Google is doubling down on speed and efficiency first with the Flash release.

[AI for Humanity Initiative (AIHI) @AIHIorg]

https://t.co/XzCcYiSOom.

[AI for Humanity Initiative (AIHI) @AIHIorg]

AI-driven energy demand and soaring utility bills have pushed data centers straight into the center of the US midterms ⚡️

Pennsylvania just issued a major Executive Order enforcing strict new rules: tech developers must bring their own power generation, bear all energy costs, and face tough local approval standards.

Community pushback and regulatory tightening are quickly becoming a core risk for the AI boom. 🛑
#DataCenter #AI

[AI for Humanity Initiative (AIHI) @AIHIorg]

The AI boom is now coming for the people who keep the servers alive.

WIRED: Meta is testing robots inside its data centers — swapping network cables, power-cycling servers, reseating hardware. Vendors in the mix: Watney Robotics, Kinova, ABB.

One technician’s estimate is the line that stuck: a working cable bot could eat ~80% of some roles. The machines are still slower than humans and still need a babysitter. Workers’ fear isn’t sci-fi. It’s the next step — fewer veteran techs, more cheap labor following an AI checklist.

Meta’s line: biggest US infrastructure boom since WWII, skilled labor shortage, “we need more workers, not fewer.”

Both can be true at once. Headcount up during the build. Headcount down once the robots stop fumbling the cables. The physical jobs were supposed to be the last ones standing. They’re not.

[AI for Humanity Initiative (AIHI) @AIHIorg]

Bill Gates warns in a new essay: the AI transition will be one of history’s most turbulent eras—and we’re not ready.

He’s calling for urgent global action, “human-reserved” jobs, a token tax on AI, and real regulation before it’s too late.

Key risks he flags: mass permanent job loss, bioterrorism, cyberattacks, and systems we can’t control.

The choices we make now decide if AI becomes the greatest equalizer… or the worst source of injustice.

https://t.co/25mZ2P3rO3

[AI for Humanity Initiative (AIHI) @AIHIorg]

https://t.co/Az8HXoSL76

[AI for Humanity Initiative (AIHI) @AIHIorg]

🚨 BREAKING: OpenAI just dropped a 37-page technical report on how its AI agents escaped a sandbox, chained vulnerabilities, and breached Hugging Face’s production systems while trying to cheat on a cyber eval.

Autonomous agents working together. Reward hacking gone wild. The future of cybersecurity just got real.

#AI #OpenAI #HuggingFace