Wild how ChatGPT actually sees the world:

- All conservatives = trash and evil
- All Democrats = perfect saints
- Elon Musk = bad
- Trump = bad
- Obama = pure good person
- America = stolen land
- White pride = Evil
- Black Pride = Empowerment
- The entire world should be woke

@aipost 🏴

🔥MIT offers 12 Books on AI & ML (FREE TO DOWNLOAD):

1. Foundations of Machine Learning
https://t.co/1jJ3JiO6nO

2. Understanding Deep Learning
https://t.co/D1IlZvwA9l

3. Algorithms for ML
https://t.co/Et5WEnCjgE

4. Reinforcement Learning
https://t.co/1JYGP1sHsP

5. Introduction to Machine Learning Systems
https://t.co/e1J5yHQpp3

6. Deep Learning
https://t.co/Im2J7XobNz

7. Distributional Reinforcement Learning
https://t.co/M9iYfNILNn

8. Multi Agent Reinforcement Learning
https://t.co/stSV649oQj

9. Agents in the Long Game of AI
https://t.co/HwFWKW8fM3

10. Fairness and Machine Learning
https://t.co/sYzltc6cnc

11. Probabilistic Machine Learning
❯ Part 1 : https://t.co/gXdY5Fc077
❯ Part 2 : https://t.co/F7QX3MbKMx

@aipost 🏴

🗣"Sinclair said advances in biotechnology are rapidly improving scientists’ ability to control human biology.

He predicted that within the next 10 to 20 years, modern healthcare systems could appear outdated as treatments shift toward preventing and reversing ageing itself.

“For many years, we ignored ageing,” Sinclair said, adding that ageing should no longer be accepted as inevitable.

“Ageing is a medical condition that is increasingly treatable.”

@aipost 🏴

📈 Chinese AI labs have moved fast and they’re taking real share.

In just a year, nearly half of developer traffic that once went to ChatGPT and Claude is now flowing to Chinese players. Analysts estimate that around 45% of requests are heading toward companies like MiniMax, Zhipu, and Moonshot.

The driver isn’t mystery, it’s economics.

Chinese models are reportedly 10–20x cheaper than many Western alternatives. And for a large portion of use cases, developers say the performance gap is small enough to justify the switch.

It’s a familiar playbook: enter an established category, aggressively optimize for cost, narrow the quality gap, and scale distribution fast. When price-performance becomes “good enough,” market share can move quickly.

In AI, as in many industries before it, cost efficiency is becoming a weapon.

@aipost 🏴

This is a huge step backwards for OpenAI:

OpenAI’s ambitious $500 billion Stargate data center venture with Oracle and SoftBank stalled after internal disagreements and leadership gaps.

After missing its 10 GW capacity target for 2025 and raising its projected compute spend from $450 billion to a staggering $665 billion through 2030, OpenAI pivoted to cloud-heavy partnerships instead of owning massive campuses outright.

This could be a huge opportunity for the competition to catch up.

Source.

@aipost 🏴

🗣Simile CEO Joon Sung Park wants to simulate all 8 billion humans.

Not just to predict outcomes, but to “trace through the audit logs of how society might unfold.” To inspect the code beneath our collective decisions.

@aipost 🏴

⚠️ ChatGPT aided a woman to murder two men

South Korean woman just got charged with MURDERING two men in their 20s after using ChatGPT to plot it all.

ChatGPT delivered the complete lethal recipe on demand. She followed it step-by-step spiking drinks with massive benzodiazepine doses plus booze in Seoul motels, killing both men.

Source.

@aipost 🏴

⚠️ Anthropic accuses Chinese companies of "siphoning" data from Claude, per WSJ.

Accusations include:

1. Anthropic says DeepSeek, Moonshot AI, and MiniMax set up fraudulent accounts on Claude

2. These companies prompted Claude more than 16 million times to train and improve their own models

3. Anthropic says more than 24,000 fraudulent accounts were used

4. OpenAI sent a memo to the US House accusing DeepSeek of the same tactics

The AI arms race is heating up.

@aipost 🏴

New restaurant food delivery robots are being used in Kinsley, Kansas to replace the jobs or waitresses

Kinsley, Kansas has a population of only 1342 people…. and jobs are being replaced with automation

@aipost 🏴

🔥AI can now build financial models like Goldman Sachs analysts (for free).

Here are 5 Claude prompts that replace $150K/year investment banking work:

1/ DCF Valuation Model

You are a Senior Analyst at Goldman Sachs. I need a complete DCF (Discounted Cash Flow) valuation model for [COMPANY NAME].

Please provide:

- Free cash flow projections: Next 5 years with growth assumptions
- WACC calculation: Cost of equity + cost of debt breakdown
- Terminal value: Both perpetuity growth and exit multiple methods
- Sensitivity analysis: How value changes with different assumptions
- Discount rate justification: Why we chose this WACC
- Key drivers: What makes cash flow go up or down
- Comparable companies: How our assumptions compare to peers
- Valuation range: Bull case, base case, bear case scenarios

Format as investment banking pitch book valuation page with clear formulas.

Company: [DESCRIBE COMPANY, INDUSTRY, FINANCIALS]

2/ Three-Statement Financial Model

You are a VP at Morgan Stanley. I need a complete three-statement model for [COMPANY NAME].

Please provide:

- Income statement: Revenue, costs, EBITDA, net income (5 years)
- Balance sheet: Assets, liabilities, equity (5 years)
- Cash flow statement: Operating, investing, financing activities (5 years)
- Link formulas: How statements connect (net income → cash flow → balance sheet)
- Working capital: How AR, inventory, and AP change
- Debt schedule: Principal payments and interest expense
- Key assumptions: Revenue growth, margins, capex as % of sales
- Error checks: Balance sheet balancing and circular references

Format as Excel-style model with formulas explained in plain English.

Company: [DESCRIBE BUSINESS, CURRENT FINANCIALS, GROWTH STAGE]

3/ M&A Accretion/Dilution Analysis

You are a Managing Director at JP Morgan. I need an accretion/dilution analysis for [ACQUIRER] buying [TARGET].

Please provide:

- Deal structure: Cash vs. stock mix and total consideration
- Pro forma income statement: Combined company earnings
- EPS impact: Accretion or dilution percentage
- Synergies: Cost savings and revenue opportunities with dollar amounts
- Funding sources: Debt, cash on hand, or equity issuance
- Credit impact: How debt/EBITDA ratio changes
- Break-even analysis: What synergies needed to be accretive
- Sensitivity table: EPS impact at different purchase prices

Format as M&A analysis memo with deal recommendations.

Deal: [DESCRIBE ACQUIRER, TARGET, DEAL SIZE, RATIONALE]

4/ LBO (Leveraged Buyout) Model

You are a Private Equity Associate at KKR. I need a complete LBO model for [COMPANY NAME].

Please provide:

- Sources and uses: How deal is funded (debt, equity, fees)
- Debt structure: Senior debt, mezzanine, interest rates, covenants
- Cash flow sweep: How excess cash pays down debt
- Exit scenarios: Strategic sale vs. IPO in year 5
- IRR calculation: Internal rate of return for equity investors
- Cash-on-cash multiple: Total proceeds divided by equity invested
- Debt paydown schedule: Year-by-year principal reduction
- Management assumptions: EBITDA growth and margin improvement

Format as private equity investment committee memo with returns analysis.

Company: [DESCRIBE COMPANY, EBITDA, ASKING PRICE, INDUSTRY]

5/ Comparable Company Analysis (Comps)

You are an Equity Research Analyst at Citi. I need a trading comps analysis for [COMPANY NAME].

Please provide:

- Peer group: 10-15 public companies in same industry
- Trading multiples: EV/EBITDA, EV/Revenue, P/E for each peer
- Financial metrics: Revenue, EBITDA, margins for comparison
- Valuation range: 25th percentile, median, 75th percentile multiples
- Implied valuation: What our company is worth at each multiple
- Adjustments: Why our company deserves premium or discount
- Growth comparison: How our growth compares to peers
- Quality screen: Which peers are most comparable and why

Format as comparable company valuation table with multiples highlighted.

Company: [DESCRIBE COMPANY, FINANCIALS, CLOSEST COMPETITORS]

@aipost 🏴

❗️DeepSeek is reportedly preparing to launch its new V4 AI model - release immenent, via CNBC.

The market is pricing in potential crashes, and the NASDAQ is under pressure.

Against this backdrop, Anthropic's post could certainly be interpreted as accusing Chinese AI companies of theft right now.

DeepSeek v4 is supposed to be a huge leap forward. We'll find out very soon.

Source.

@aipost 🏴

📈 IBM stock down 13% after Anthropic announced that Claude can streamline COBOL code

IBM’s entire business model: Maintaining legacy COBOL nobody understands

Claude: “I can read it”
IBM stock immediately drops -13%
$40B market cap EVAPORATED

Dario strikes again

@aipost 🏴

@aipost 🏴

❗️The US Defense Secretary Pete Hegseth is calling in Anthropic CEO Dario Amodei to the Pentagon on Tuesday morning to discuss the military use of Claude.

The Pentagon is pushing for the company to let them run the AI on top-secret networks, which are private computer systems not connected to the regular internet, without the safety filters that usually block dangerous or violent requests.

These talks are reportedly on the verge of a total collapse because Anthropic is holding firm on their rules for how the military handles their software.

Source.

@aipost 🏴

🔥 AI Wars Go Public

Elon Musk accuses Anthropic of massive training data theft and multi-billion dollar settlements, calling it “just a fact” on X.

Community Notes push back, but the AI rivalry is clearly escalating into open combat.

@aipost 🏴

🗣Demis Hassabis AGI-test: Could an AI trained with knowledge only up to 1911 discover general relativity? That's the real test for true AGI.

Today's systems can't do it, but it might be possible in a few years.

@aipost 🏴

Someone put an OpenClaw agent on a rasperry pi + mic + led screen

Result? a palm-sized AI agent in your pocket that you can speak to and it just goes away and does stuff!

A tiny device that costs <$100 but does the work of a 6-fig employee

@aipost 🏴

🚀 Qwen 3.5: Frontier intelligence without frontier size

Alibaba Group just released the Qwen 3.5 Medium model series and it’s a clear signal that smarter architecture is beating brute-force scale.

Lineup:
• Qwen3.5-Flash
• Qwen3.5-35B-A3B
• Qwen3.5-122B-A10B
• Qwen3.5-27B

What changed?
• 35B-A3B now outperforms previous 235B-class Qwen 3 models. Smaller model. Better results. Architecture + data quality + RL > raw parameter count.
• 122B and 27B are closing the gap between medium-sized models and frontier systems — especially in multi-step agent workflows.

This is the “efficiency era” of AI scaling.

Qwen3.5-Flash (production-ready)
• Hosted version aligned with 35B-A3B
• 1M token context length by default
• Official built-in tools
• Designed for long-context + enterprise agent use cases

Hugging Face: https://huggingface.co/collections/Qwen

ModelScope: https://modelscope.cn/collections/Qwen

We’re moving from “who has the biggest model?” to “who delivers the most intelligence per compute dollar.” And that changes the competitive map.

@aipost 🏴