'Don't let LLM's success cloud your judgment': Tech CTO shares hard-hitting AI truths for businesses
More Than Just LLMs
Let Curiosity Lead, Not Cost
Why This Matters Now
You Might Also Like:
GitHub CEO calls out AI panic, explains why the idea of coding skills becoming obsolete is 'mistaken'
Speaking at the recent MIT Technology Review EmTech AI Conference, Akamai CTO Robert Blumofe offered a refreshingly grounded perspective on how enterprises can break free from the relentless " AI hype cycle "—a pattern where curiosity turns to FOMO, and hastily adopted AI solutions lead to disappointment. His four-point roadmap, shaped by Akamai's own AI journey, serves as a crucial reality check in a world increasingly driven by artificial intelligence.Blumofe, who also holds a PhD in computer science from MIT, described a familiar trap that many organizations are falling into. 'That's the chain: AI success, theater, FOMO, and some form of failure,' he said during his talk. Businesses, in their rush to appear cutting-edge, mistake early-stage use cases for scalable solutions—plunging into costly and ineffective AI deployments.And this problem isn't niche. According to a Pew Research study cited in his address, only 1 in 6 U.S. workers currently use AI at work, revealing a stark gap between AI's perceived and practical utility. 'Most jobs at this point can benefit from AI,' said Blumofe. 'It's a matter of which tasks can most benefit, and how, using which form of AI.'Blumofe urged companies to look beyond the fascination with large language models . While LLMs like ChatGPT have demonstrated remarkable versatility—from email classification to customer support—they're not the silver bullet for every enterprise challenge.'In many ways, an LLM is a ridiculously expensive way to solve certain problems,' he noted, pointing to Akamai's use of purpose-built models in cybersecurity threat detection. Models like these, he argued, offer more efficiency and relevance than a trillion-parameter generalist.His advice? Think smaller and sharper. LLMs are just one tool in a vast AI toolkit. Symbolic AI, deep learning, and ensemble models can be better suited for tasks that require precision, logic, and domain specificity.Akamai's approach to fostering AI adoption is democratic: let employees experiment. The company built an internal AI sandbox, giving teams the freedom to play, build, and discover practical applications on their own terms. While the setup may test IT infrastructure limits, Blumofe insists the freedom sparks innovation. 'I feel no need to evaluate each use case,' he said.And when asked about companies that require hiring managers to prove AI can't do a job before hiring a human, Blumofe didn't mince words: 'That's getting the tail before the dog.' The question shouldn't be, 'Why not AI?' but 'What's the right tool for the problem at hand?'Blumofe's caution comes at a pivotal moment in AI's evolution. As VentureBeat recently reported, major players like OpenAI, DeepMind, and Meta are collaborating to raise alarms about AI systems potentially becoming too smart—and too opaque. A recent paper on 'Chain of Thought Monitorability', endorsed by AI luminaries like Geoffrey Hinton, warns that if LLMs start thinking in ways we can't interpret, we risk losing control.That's why responsible leadership matters now more than ever. The real AI revolution won't be won by the company with the flashiest chatbot—but by the one that knows exactly when, why, and how to use it.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
10 minutes ago
- Time of India
Who is Shengjia Zhao, OpenAI co-creator and now Meta's new Superintelligence chief?
Academy Empower your mind, elevate your skills Talent poaching continues at full speed among Big Tech and artificial intelligence (AI) hyperscalers. Meta has appointed Shengjia Zhao , a prominent research scientist at OpenAI , as the head of its ambitious Superintelligence team, which Mark Zuckerberg is was a prominent research scientist at OpenAI, where he played a major role in the creation of ChatGPT , GPT-4, and several notable 'mini' models, including GPT-4.1 and has co-authored the original ChatGPT paper and is credited with starting trends in AI reasoning, particularly through his work on OpenAI's o1 has a PhD in computer science from Stanford University and holds a bachelor's degree from Tsinghua Meta, he will be the lead scientist of the Superintelligence team and will direct the lab's research agenda, working with Meta CEO Zuckerberg and chief AI officer Alexandr is entrusted with setting the strategy for Meta's foundational models, especially the Llama series, and unifying the company's efforts to compete at the frontier of foundational AI research Zhao made major foundational contributions to ChatGPT during his tenure at OpenAI. He also led key efforts in synthetic data generation used to train large language models (LLMs) like ChatGPT and its addition to his direct work on ChatGPT, Zhao helped create the o1 and o3 reasoning models, with o1 giving rise to the "chain-of-thought" prompting technique, adopted by models across the space tech giant launched the Meta Superintelligence Labs (MSL) in 2025 to supercharge its ambition of creating artificial general intelligence (AGI) and, ultimately, superintelligence. This new division will bring together Meta's foundational AI research (including Llama models), product AI teams, and next-generation model efforts under one umbrella.
Time of India
24 minutes ago
- Time of India
IPO or M&A? Why Indian promoters are choosing both
As India steps into the second half of 2025, the country's most ambitious Founders are exiting differently. IPO and M&A are no longer a binary choice. This is the dual-track exit era: When companies prepare for an IPO but keep strategic or financial M&A discussions alive. Long considered a hedge in Western markets, this approach is now an intentional and predominant playbook in India. And this is not just talk. IPO filings of over INR 1.1 lakh crore (USD ~12.7bn) in H1, and more than USD 20bn in disclosed M&A deals, including inbound interest in fintech, NBFCs , Q-commerce , and defence, speak for themselves. However, the untold story lies deep beneath the filings, where Founders are engaging in a parallel choreography of multiple processes. Explore courses from Top Institutes in Please select course: Select a Course Category Technology Operations Management Public Policy Artificial Intelligence Finance Management MBA others CXO healthcare Healthcare PGDM Cybersecurity Data Analytics MCA Leadership Product Management Design Thinking Data Science Project Management Others Degree Data Science Digital Marketing Skills you'll gain: Duration: 12 Weeks MIT xPRO CERT-MIT XPRO Building AI Prod India Starts on undefined Get Details From Backup Plan to Built-In Strategy IPO used to be the aspirational route, while M&As always seemed to be Plan B. But 2025 has replaced this thinking. Today, founders prepare for IPOs to unlock M&A premium and vice versa. What Has Changed? Firstly, regulatory uncertainty posed roadblocks at the structural level. The reverse flip, i.e., the relocation of Indian-origin startups from Delaware/Singapore back into India, is now quite routine. Several high-growth companies have demonstrated that it is not only feasible but also rewarding, particularly with Securities and Exchange Board of India (SEBI) increasingly attending to technological intricacies in IPO vetting, along with MCA/RBI setting up easements in compliance procedures. Secondly, late-stage capital has become selective, thereby forcing Founders to bake in some measures of optionality. The mere preparation of a Draft Red Herring Prospectus (DRHP), with its rigour in compliance, disclosures, and governance, forces a company to become exit-ready, however, whichever door it walks through. Thirdly, market volatility exists. What might seem like a golden IPO window in Q1 can suddenly be gone by Q3. Running a synchronised M&A process provides a fallback or, in many cases, a much speedier, cleaner route. IPO vs. M&A: What Founders Stand to Gain IPOs pave the way for long-term growth capital, boost brand visibility, and create a public currency for future deals. M&As bring speed, valuation clarity, and the potential for strategic synergies. Today, Founders are not choosing one over the other; they are leveraging both. Dual tracking is no longer a fallback; it's a forward-looking strategy to maximise value, retain leadership, and shape the future on their terms. Real-Time Instances: How India Is Playing The Dual Game A leading fintech unicorn filed its DRHP to raise INR 2,600 crores, post its reverse-flip back to India in April 2025. No M&A deal is prima facie announced; however, indications suggest that the company is setting IPO valuation as the pricing anchor for strategic acquirers. Meanwhile, a prominent quick commerce player has quietly advanced its dual-track agenda. After re-domiciling to India in 2024, its path to IPO was always expected. However, multiple suitors are reportedly circling, and the company seems to be using M&A interest as a pricing test before officially filing its DRHP. In the industrial space, a major building materials company has taken a textbook dual-track approach. While it has SEBI's nod for an INR 4,000 crore IPO, it is also actively exploring inorganic expansion via distressed asset acquisitions under India's Insolvency and Bankruptcy Code (IBC) framework. This two-pronged play supports growth and enhances the company's appeal to both public investors and strategic buyers. These are no longer isolated cases, but structural changes affecting the exit planning side, as well as the way Indian companies view liquidity strategy in comparison with timing, readiness, and leverage. Dual-Track Being Execution-Heavy Is Worth It The true dual-track process is by no means a passive hedge, but an execution-intensive strategy. Companies must maintain two parallel data rooms - one for IPO diligence and the other for private buyers. The narrative must be tailored differently - growth, governance and scalability for public markets; synergies and market consolidation for strategic buyers. Advisors such as bankers, lawyers, and consultants need to be coordinated to run parallel but distinct processes. Public disclosures like DRHP filings must be timed with precision. They can be used tactically to spark or strengthen M&A conversations, creating price tension across both tracks. What H2 2025 Will Bring The IPO calendar for H2 2025 is packed, with several large companies across infrastructure, fintech, and consumer sectors preparing to test public market appetite. Meanwhile, strategists sit on record dry powder. Global majors in defence, payments, and consumer internet are scouring India not for early-stage bets but for exit-ready, governance-strong assets. Pre-IPO dealmaking is likely to accelerate, particularly for companies in the INR 500–2,000 crore revenue band, where public market appetite may remain sector-sensitive or ambiguous. Dual-track exits will also expand beyond tech into new verticals such as renewables, EV infrastructure, healthcare delivery, and logistics. Interestingly, the DRHP itself is gaining a new role, less as a declaration of IPO intent, more as a pricing discovery tool. Increasingly, founders and boards are using it as the new term sheet, signalling readiness and setting valuation benchmarks that strategic buyers must react to. The Bigger Picture We believe the rise of dual-track exits marks more than just a tactical shift; it signals a maturing of India's private capital ecosystem. This is where Founder ambition meets institutional discipline. Companies are no longer reactive to market sentiment; they're proactively shaping outcomes with optionality built into every decision. Dual tracking is not about indecision, but intelligent design. For founders, it means greater control over timing and valuation. For investors, it offers more flexible exit pathways. And for India Inc., it points to a new era of dealmaking, one that blends global capital sensibilities with India's appetite for innovation and growth. (The author is , Partner, Deal Value Creation Services, BDO India)
Hans India
40 minutes ago
- Hans India
OpenAI CEO Sam Altman Warns: ChatGPT Conversations Not Legally Confidential
As artificial intelligence tools become more entwined with daily life, many users have turned to platforms like ChatGPT for emotional guidance, life coaching, and even informal therapy. However, OpenAI CEO Sam Altman has now raised an important red flag: your private conversations with ChatGPT might not be as private as you think. In a candid discussion on This Past Weekend, a popular YouTube podcast hosted by comedian Theo Von, Altman shared that conversations users have with ChatGPT—no matter how sensitive or personal—are not protected by legal confidentiality. This means that, under certain circumstances such as a lawsuit, these interactions could be made public. 'People talk about the most personal sh*t in their lives to ChatGPT,' Altman remarked. 'People use it – young people, especially, use it – as a therapist, a life coach; having these relationship problems and [asking] what should I do? And right now, if you talk to a therapist or a lawyer or a doctor about those problems, there's legal privilege for it. There's doctor-patient confidentiality, there's legal confidentiality, whatever. And we haven't figured that out yet for when you talk to ChatGPT.' This acknowledgment has sparked fresh concerns among users and privacy advocates, especially considering how deeply integrated these AI tools have become in everyday emotional and mental wellness routines. Altman emphasized the urgency of developing policies that extend similar confidentiality protections to AI interactions. 'So, if you go talk to ChatGPT about your most sensitive stuff and then there's like a lawsuit or whatever, we could be required to produce that, and I think that's very screwed up,' he said. This revelation underscores a major gap in current legal frameworks, especially as AI is rapidly being adopted across sectors. Unlike end-to-end encrypted platforms like WhatsApp or Signal, which are designed to keep user communications private and unreadable by third parties, ChatGPT's infrastructure allows OpenAI to access conversations. This access, according to the company, helps improve the model and monitor for potential misuse. Although OpenAI has stated that it deletes free-tier user chats within 30 days, the company has also noted that some interactions may be retained for legal or safety-related purposes. The issue becomes even more relevant as OpenAI faces a high-profile lawsuit filed by The New York Times, which has compelled the company to preserve user conversations—excluding those from enterprise accounts—for legal scrutiny. As AI becomes more trusted with intimate parts of our lives, Altman's remarks make it clear that the industry has yet to address the need for legal and ethical protections that match those of human professionals. Until then, users are urged to exercise caution and understand the limitations of confidentiality when engaging with AI chatbots for personal or mental health matters.
