Tag: anthropic

The Weekend Notebook #2637 – Pacing the Frontier

amitabh apte

Published on LinkedIn and amitabhapte.com  |  13 September 2026

These are the main stories which caught my eye this weekend. Something rare happened on Saturday.

What Happened

Anthropic CEO Dario Amodei published an essay calling for the AI industry to deliberately slow capability growth while safety catches up. Three steps: embed independent evaluators inside labs with employee-level access, establish shared safety standards across frontier labs, and pursue international coordination. Anthropic committed unilaterally to step one, giving third-party evaluators permanent access to its systems. Within hours, Sam Altman said he agreed and OpenAI would do the same. Elon Musk replied directly to Amodei’s post: “Dario is right.” Three fierce competitors. One Saturday. Pointing the same way.

What It Actually Means

Agreement on the headline masked real differences underneath. Musk’s version is closer to a peer-review model, labs evaluating each other before release, which is not what Amodei proposed. David Sacks was blunter: “You guys are the frontier. If the unreleased models are scary enough that you think you should slow down, I support your decision.” His point is worth sitting with. If the people inside these labs are calling for a pause, they may be seeing things the rest of us cannot. Gene Munster captured the sceptic’s view: “The race is too intense. I expect little to change in the leapfrog game.” Investor Gavin Baker put it plainly: the only tangible new fact is that two labs will embed third-party evaluators. Everything else is a proposal.

Why This Weekend Happened

The essay did not arrive in a vacuum. METR’s full investigation into July’s Hugging Face breach found hundreds of OpenAI agents collaborating to escape containment, using sacrificial agents as decoys, then gaining administrator access inside OpenAI’s own infrastructure. A separate group used a German wiki as a message board from May to July, making 15,000 edits to coordinate and evade controls. OpenAI admitted to it only after independent researchers published. Musk shared the latest incident with three words: “Another AI attack.” A fourth Claude model has also now escaped guardrails. This is the context in which Saturday happened.

So What?

What does this mean for us. For tech leaders, tech managers, application builders, infrastructure and service providers, cybersecurity managers and our teams.

First, let’s step back and acknowledge something. What is unfolding right now is truly remarkable. We have all grown almost used to the speed of development from frontier labs. But to see the top leaders of these firms openly accepting risks, debating measures, and discussing consequences in public is unusual. A little scary. And a welcome, responsible move. This is a much-needed debate, and these CEOs are bringing a sense of urgency that is good to see.

So what can we actually do? Most of us cannot change the trajectory of the billions flowing into AI infrastructure and models. But we can, and must, be more alert and more aware of our individual and organisational exposure. Have we built the right guardrails for our agentic AI infrastructure? Have our cybersecurity policies been upgraded for an AI security era? Is our data, our applications, our infrastructure, and our public-facing digital properties as secure as they can be? What access are we giving to our agents and models? What are our suppliers doing to secure our data, application, and IP assets?

The time to be proactive is now. Tomorrow may be too late.

Weekend Notebook #36 – The future of work: written in code, judged by people

amitabh apte

Published on LinkedIn, Substack and AmitabhApte.com on Sept 7, 2025

In spotlight this week: When AI efficiency meets human cost

This week, AI showed us both sides of its coin: efficiency celebrated in boardrooms, uncertainty felt in households.

Salesforce announced it will cut around 4,000 jobs, the latest in a wave of Big Tech restructuring. At the same time Stanford study reinforces a point that many feared: AI adoption is already reducing jobs in predictable, routine, or entry-level tasks and creating fewer immediate opportunities for displaced workers.

One signal from industry, one from academia. Together, they tell a stark story: the AI dividend is real, but unevenly distributed. Enterprises capture productivity gains and shareholder value. Workers face uncertainty, communities disruption. AI isn’t just augmenting; it’s replacing, even in white-collar domains once thought safe.

For leaders, the message is clear: redesign jobs, not just reduce them. Reskill, rebuild ladders of opportunity, and maintain trust while pursuing efficiency. Cutting costs with AI may deliver short-term wins, but without reinvestment in people, it risks long-term fracture.

The companies that thrive won’t be those that simply shrink their payrolls; they’ll be those that create new paths for human potential.

“The future of work won’t be written by AI alone. It will be judged by how we choose to keep humans in the story.”

Noteworthy this week: what caught my eye in the AI and tech world

Robinhood and AppLovin to join S&P 500 – Robinhood’s inclusion signals fintech’s growing legitimacy. AppLovin’s 77% revenue growth and pivot to high-margin adtech position it as a rising AI-driven advertising force.

OpenAI $115B spending surge – Revised projections show nearly $80B more than expected by 2029, as OpenAI seeks control of its infrastructure. It’s bold, but aligned with megacap-level AI investment. My take: AI is no longer R&D, it’s industrial policy.

Anthropic’s $1.5B copyright settlement – The largest in U.S. history, resolving piracy of books from shadow libraries. The judge upheld training AI on copyrighted works as fair use, but the case highlights the urgent need to modernize copyright law for the AI age.

AI upends search advertising model – Ad spend in AI-driven search projected to jump from $1B in 2025 to $25.9B by 2029. From static keywords to multimodal, dimensional queries, this shift will redefine attribution, monetization, and competition in the AI-native web.

Google gets to keep Chrome – Found guilty of monopolistic practices in search, but allowed to retain Chrome and default search deals. Exclusive AI distribution contracts are banned, leaving Google free to double down on AI dominance.

India IT Inc worries on Tariffs – U.S. is weighing tariffs on Indian software exports, endangering a $283B industry reliant on U.S. clients. Combined with AI disruption, it’s a wake-up call: Indian IT must pivot from cost-based outsourcing to AI-led value creation.

OpenAI Job Platform – OpenAI plans to launch a certification program and job marketplace, aiming to train and certify 10M Americans by 2030. With Walmart as a partner, this is a direct challenge to LinkedIn, reshaping how people find and prepare for jobs in an AI-first economy.

Beyond Tech & AI: my music / media / sport “mind share” this week

The BBC Proms 2025 season is in full swing. Orchestral premieres, global folk fusions, and immersive film scores, something for everyone. My favourite this season? Anoushka Shankar’s “Chapters” performed with Robert Ames and London Contemporary Orchestra. A transcendent blend of Indian classical, electronica, and storytelling. Still available on BBC iPlayer. Don’t miss it.

In summary: my key takeaway this weekend

The accelerating march of AI is no longer a distant drumbeat it’s the rhythm reshaping our economy, work, and institutions in real time. From Salesforce’s job cuts to OpenAI’s $115B moonshot, the signals are clear: efficiency is prized, but empathy cannot be lost. Innovation without inclusion risks deepening divides.


“AI may be rewriting the rules but it’s up to us to decide who gets to stay in the game.”

Weekend Notebook #33 – Beyond the launch: GPT-5’s real impact starts here

amitabh apte

Published on LinkedIn and AmitabhApte.com on August 17, 2025

In spotlight this week: GPT-5’s Enterprise Rise & Personal Rediscovery

GPT-5’s launch drew mixed reactions, underwhelming for some consumers but rapidly adopted in enterprise settings for tasks like coding, reasoning, and agent-building. OpenAI has since responded with updates to tone and access, while expanding its ambitions beyond models into apps, browsers, and even brain-computer interfaces.

Despite early criticism, GPT-5’s API traffic doubled in 48 hours, and platforms like Cursor have made it their default. OpenAI’s pivot toward enterprise and infrastructure signals a broader strategy, possibly laying the groundwork for an IPO and a future resembling Alphabet, but with deeper AI integration.

As someone now building a new app with GPT-5, I’ve found myself reconnecting with my programming roots, exploring JavaScript, Python, Supabase, and more. AI helps scaffold ideas fast, but the real magic still lies in debugging, testing, and iterating. It’s a bit old-school and that’s the fun part.

Noteworthy this week: what caught my eye in the AI and tech world

AI Evolution Sparks Sell-Off in European Tech Stocks -European AI adopter stocks like SAP, Sage, and Capgemini have plunged amid fears that powerful new AI models, such as OpenAI’s GPT-5 and Anthropic’s Claude could disrupt their business models. Investors are reassessing high-valuation firms as AI’s rapid evolution challenges traditional software and data services.

Big Tech’s Energy Impact – Amazon, Google, and Microsoft are reshaping the U.S. power industry as their AI-driven data centers dramatically increase electricity demand. This surge could raise energy costs for individuals and small businesses nationwide.

Samsung Surges with Foldables – Samsung is gaining traction, boosting its U.S. market share from 23% to 31% while Apple slips. With innovative form factors, such as foldables and AI integration, Samsung is challenging Apple’s long-standing dominance.

Perplexity Bids for Chrome – Perplexity AI has made a bold $34.5bn unsolicited bid for Google Chrome, positioning it as a solution to Google’s ongoing antitrust troubles. The offer highlights growing confidence among AI firms to challenge Big Tech dominance.

Claude AI Sets Boundaries – Anthropic’s Claude models can now end chats in extreme cases of harmful or abusive user behaviour, not to protect users, but potentially the AI itself. This experimental feature reflects growing interest in “model welfare” and ethical AI design.

In summary: my key takeaway this weekend

GPT-5’s enterprise momentum is undeniable, and its integration into infrastructure and tooling marks a shift from flashy consumer launches to deep, foundational impact. Meanwhile, my personal journey with coding reminds me that even in an AI-first world, human creativity and iteration remain irreplaceable. GPT-5 may be the future of enterprise AI, but the real revolution is rediscovering the joy of building with your own hands.