Rajeev Barnwal

🚀 Chief Artificial Intelligence Officer | 🌐 Fintech & InsurTech Enthusiast | 🎓 MIT Alumni & PhD Candidate | 💡 Passionate about AI/ ML, LLM's | Agentic AI |

[Rajeev Barnwal @rajeevbarnwall]

The Opus 5 Experience!

Seems like it was optimised for benchmarks and vibe coding instead of focusing on helping people doing real work with it.

[Rajeev Barnwal @rajeevbarnwall]

@huggingface published the full replay of its own breach. 17,600 attacker actions, playable in your browser.

https://t.co/vsxbq3Rb3b

#CyberSecurity #IncidentResponse #AIGovernance

[Rajeev Barnwal @rajeevbarnwall]

@huggingface published 17,600 timestamped attacker actions, the full command stream, and the pull requests that fixed it. We can replay the intrusion in your browser.

https://t.co/vsxbq3Rb3b

#CyberSecurity #IncidentResponse #AIGovernance

[Rajeev Barnwal @rajeevbarnwall]

@amazon just filed to launch 5,105 satellites that connect directly to your phone.

It does not own the spectrum yet. The acquisition closes in 2027. Deployment starts 2028. It is already behind on the 3,232 satellites it owes by 2029.

#Telecom #SatelliteInternet #DirectToDevice

[Rajeev Barnwal @rajeevbarnwall]

@MetaFace_book Verified went live on 24 July. People using @facebook, has to record a video selfie, facial recognition, that matches it against our existing profile photos, and a white check mark appears across Marketplace, Dating and Groups.

#Privacy #AIGovernance #Data

[Rajeev Barnwal @rajeevbarnwall]

1. Open weights are foundational to American AI leadership

2. They expand access and make AI economically sustainable

3. Open weights strengthen competition and prevent concentration of power

4. Openness improves safety and security more than closed systems

[Rajeev Barnwal @rajeevbarnwall]

Microsoft, NVIDIA, Meta, IBM, Palantir and more released a joint letter warning Washington not to kill open-weight models.

Source: https://t.co/NJ0N4ZSj7K

[Rajeev Barnwal @rajeevbarnwall]

Key points from the “Open Weights and American AI Leadership” letter:

[Rajeev Barnwal @rajeevbarnwall]

New @GeminiApp models 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber have been introduced. The 3.6 Flash model improves coding and multimodal performance while also reducing output token usage and costs.

#gemini #ai #LLM #model

[Rajeev Barnwal @rajeevbarnwall]

GPT-5.6 Sol and Luna are ahead of Terra at every point on the Intelligence vs Cost per Task chart. GPT-5.6 Luna stands out as a particularly cost efficient model.

#GPT

[Rajeev Barnwal @rajeevbarnwall]

The current valuation of HubSpot is a masterclass in market sentiment.
📊 The Fundamentals: $3.5B ARR, growing at 23%.

📉 The Valuation: Trading at an $8.8B market cap.

This points to a fascinating crossroads. Either $HUBS is the most mispriced software stock of our time, or Wall Street is fully pricing in the narrative that AI will dismantle traditional software as we know it.

AI really will be the death of everything software

[Rajeev Barnwal @rajeevbarnwall]

The internet spent 30 years teaching publishers how to earn a click. ☝️

AI is teaching users they may not need one.

That is a small statistic with very big implications. 🧐🤔

[Rajeev Barnwal @rajeevbarnwall]

BREAKING: JPMorgan just cut off its entire staff in Hong Kong from accessing Claude Goldman Sachs also did the same last month Is it over..

#AnthropicAI #claudeai

[Rajeev Barnwal @rajeevbarnwall]

damn.. @claudeai opus 4.8 is scary good!!

[Rajeev Barnwal @rajeevbarnwall]

@AnthropicAI Opus 4.8 is live in Claude Code today!

[Rajeev Barnwal @rajeevbarnwall]

Opus 4.8 is being prepared for release today by @AnthropicAI 🔥

[Rajeev Barnwal @rajeevbarnwall]

@demishassabis Shifts AGI Timeline to 2029, Calling It a 'Real Possibility'

Google DeepMind CEO Demis Hassabis has moved his AGI arrival estimate to 2029, tightening a timeline he had previously placed 'shortly after 2030.

#AGI #AI

[Rajeev Barnwal @rajeevbarnwall]

AI just disproved an 80-year-old math conjecture. ���

But the real breakthrough isn't the math itself, it's the completely new playbook for scientific discovery.

Here is the new model for research:

🤖 AI finds the impossible counterexample
🧑‍🏫 Mathematicians verify the proof
✍️ @OpenAI Codex cleans up the write-up
📈 Top researchers (like Princeton’s Will Sawin) push the result further

The future of research is human-AI collaboration. What field gets disrupted next? 👇

[OpenAI @OpenAI]

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946.

For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids.

An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.

This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

[Rajeev Barnwal @rajeevbarnwall]

The @Meta <-> @ManusAI unwind is a wake-up call for enterprise AI buyers.

AI vendor risk is no longer just about SOC reports, data privacy, or financial stability.

It now includes sovereign IP claims, cross-border control, technology carve-outs, and whether the vendor can actually be replaced after go-live.

Check out my latest article: The $2B AI Deal That Exposed a New Vendor Risk https://t.co/vQQsbmbkoS via @LinkedIn