Pages

Showing posts with label Mark Cuban. Show all posts
Showing posts with label Mark Cuban. Show all posts

Monday, January 05, 2026

Are Today’s AI Giants Tomorrow’s Relics?



Are Today’s AI Giants Tomorrow’s Relics?

What Tech History’s Metaphors Reveal About the Future of Generative AI

Every technological revolution carries a paradox: the very innovations that define an era often plant the seeds of their own obsolescence. Generative AI—and large language models (LLMs) in particular—now sit squarely inside that paradox. To understand where this field may be heading, it helps to borrow metaphors from tech history. Not because history repeats exactly, but because it rhymes loudly enough to hear the warning bells.

The metaphors comparing today’s LLMs to extinct or absorbed tech giants—DEC, Cray, Wang, Sun Microsystems, Compaq—are not casual provocations. They are diagnostic tools. They force us to ask whether OpenAI, Google, Anthropic, and their peers are building the Unix of the AI era… or the Radio Shack.

LLMs as the DEC and Sun of Tomorrow

The analogy that today’s dominant LLMs may one day resemble Digital Equipment Corporation or Sun Microsystems is provocative—and partially accurate. These firms once defined entire categories. DEC ruled minicomputers. Sun powered the internet’s early backbone. Cray was synonymous with supercomputing. Yet all were eventually eclipsed by structural shifts: microprocessors, commoditization, open standards, and new economic models.

The strength of this metaphor lies in its emphasis on disruption. LLMs today dominate not because they are perfect, but because they arrived first at scale. History suggests that first movers in foundational technology layers are rarely the last movers.

But there is also a weakness here. LLMs are not hardware. Their decline, if it comes, will not look like shuttered factories or liquidation sales. It may look more subtle: open-sourcing, fine-tuning, distillation, absorption into platforms, or quiet relegation to infrastructure. Sun “died,” but Java did not. DEC vanished, but its ideas live on in modern computing abstractions. Similarly, even if today’s AI leaders lose dominance, their models, techniques, and datasets may become permanent scaffolding.

The real lesson from these companies is not extinction—it is loss of centrality.

LLMs as Operating Systems: The AI Interface Layer

A second metaphor frames LLMs as operating systems—CP/M, DOS, Windows, Unix—sitting between raw compute and human intent. This is compelling. Increasingly, LLMs are not applications; they are the interface through which applications are built, queried, and orchestrated. APIs, plugins, agents, copilots—these are not unlike the application ecosystems that flourished atop operating systems.

Unix is the cautionary counterexample. Built decades ago, it still underpins modern systems because it was modular, portable, and conceptually clean. If LLMs become similarly foundational, they could persist far longer than skeptics expect.

Yet the analogy breaks down in important ways. Operating systems are deterministic. LLMs are probabilistic. OS kernels do not hallucinate. They do not require constant retraining. They do not consume city-sized power plants to function. Moreover, OS dominance emerged through standards (POSIX), while AI today is fragmenting along proprietary data, compute, and alignment philosophies.

If LLMs are operating systems, they are unstable ones—still evolving while already being depended upon.

LLMs as the TCP/IP of the AI Era

The most radical metaphor flips everything upside down: what if LLMs are not the visible interface at all, but the invisible protocol layer—more TCP/IP than Windows?

TCP/IP transformed the world precisely because it disappeared. Users never “used” it; they benefited from what it enabled. Email, the web, streaming, cloud computing—all rode quietly on its back.

There is a plausible future in which LLMs follow the same path. They become cheap, ubiquitous, and embedded—powering reasoning engines, agents, simulations, and multimodal systems without ever being directly experienced as chatbots. In this world, the value migrates upward to orchestration, integration, and experience.

The obstacle is economic. TCP/IP succeeded because it was open, minimal, and cheap. Today’s frontier LLMs are closed, massive, and expensive. For them to become invisible infrastructure, commoditization must occur—and commoditization is kryptonite to moats. If this happens, today’s leaders may survive, but their valuations, power, and narrative dominance may not.

The Missing Ingredient: “World View”

Beneath all these metaphors lies a deeper claim: current LLMs do not truly understand the world.

They are extraordinary pattern recognizers, but brittle reasoners. They write eloquently while occasionally failing kindergarten physics. This critique is not fringe. Researchers like Yann LeCun have repeatedly argued that autoregressive language models lack internal world models—causal, physical, and commonsense representations of reality.

The next major leap may not be a larger model, but a deeper one: systems that integrate perception, memory, causality, simulation, and planning. Video-trained models, embodied AI, neurosymbolic hybrids, and agentic systems are early signals of this shift. Efforts like Cyc failed in the past, but the difference now is scale, compute, and data diversity.

Ironically, incumbents may be best positioned to build “World View” systems—yet least incentivized to disrupt their own successful abstractions.

The AI Innovator’s Dilemma

This brings us to a classic trap: Clayton Christensen’s Innovator’s Dilemma.

Today’s leaders must keep their current users happy while betting on a future that may make their current products obsolete. Google must protect search revenue while reinventing information access. OpenAI must balance API stability with frontier breakthroughs. Every incremental improvement risks delaying a necessary paradigm shift.

Unlike Nokia, however, today’s AI leaders have a possible escape hatch: gradual evolution. Plugins, agents, multimodality, fine-tuning, and tool use allow for incremental transition. The real dilemma may not be technical, but economic—how to fund exponentially rising R&D costs while prices fall and competition intensifies.

Burning Cash While Chasing the Unknown

The financial strain is real. Training frontier models already costs hundreds of millions of dollars. Future systems—especially those aspiring to world models or AGI-level reasoning—could require orders of magnitude more compute. Sam Altman has openly speculated about trillion-dollar-scale infrastructure.

Efficiency gains like Mixture-of-Experts, quantization, and distillation may bend the cost curve, but black swans loom: energy constraints, regulation, geopolitical shocks, or open-source breakthroughs that collapse pricing power overnight.

In this environment, it is not irrational to imagine once-dominant AI firms becoming the Compaqs of the 2030s—important, remembered, but no longer central.

Commoditization, Open Source, and the Linux Scenario

If incumbents stumble, open-source models may play the role Linux played against Windows: less polished, more fragmented, but relentlessly adaptive. Models like Llama, Mistral, and their derivatives already outperform closed systems in niche domains.

This future is messy. Interoperability suffers. Standards lag. But innovation accelerates. The AI stack may begin to resemble the internet itself: layered, chaotic, and impossible to control from the top.

A Final Counterpoint: Adaptation Is Underrated

History also offers a hopeful lesson. Microsoft survived DOS, Windows, the web, mobile, and cloud—not by predicting perfectly, but by adapting relentlessly. IBM shed hardware to become a services giant. Decline is not destiny.

If today’s AI leaders successfully integrate “World View” systems—through acquisition, partnership, or internal reinvention—they may dominate longer than skeptics expect. But the center of gravity will shift. The winners of the next decade may not be model builders at all, but integrators: companies that combine AI with devices, platforms, trust, and distribution.

In the end, the metaphors are not warnings of collapse—they are reminders of humility. In technology, nothing stays at the center forever. Even the most powerful abstractions eventually fade into infrastructure, legend, or footnote.

The only certainty is this: the age of simple chatbots is already ending. Whatever comes next will not just talk about the world—it will model it.



क्या आज के एआई दिग्गज कल के अवशेष बन जाएंगे?

तकनीकी इतिहास के रूपक हमें जनरेटिव एआई के भविष्य के बारे में क्या बताते हैं

हर तकनीकी क्रांति अपने भीतर एक विरोधाभास समेटे होती है: जो नवाचार किसी युग को परिभाषित करते हैं, वही अक्सर उसके अंत के बीज भी बो देते हैं। जनरेटिव एआई—और विशेष रूप से बड़े भाषा मॉडल (LLMs)—आज इसी विरोधाभास के केंद्र में खड़े हैं। यह समझने के लिए कि यह क्षेत्र किस दिशा में जा सकता है, तकनीकी इतिहास से लिए गए रूपकों की ओर देखना उपयोगी है। इसलिए नहीं कि इतिहास स्वयं को हूबहू दोहराता है, बल्कि इसलिए कि वह इतनी जोर से तुकबंदी करता है कि चेतावनी की घंटियाँ साफ सुनाई देने लगती हैं।

आज के प्रमुख LLMs की तुलना DEC, Cray, Wang, Sun Microsystems और Compaq जैसी विलुप्त या निगले गए तकनीकी दिग्गज कंपनियों से करना मात्र उकसावा नहीं है। यह एक निदानात्मक उपकरण है। यह हमें यह पूछने पर मजबूर करता है कि क्या OpenAI, Google, Anthropic और उनके समकक्ष एआई युग का Unix बना रहे हैं—या Radio Shack।


कल के DEC और Sun के रूप में LLMs

यह रूपक कि आज के प्रमुख LLMs एक दिन Digital Equipment Corporation (DEC) या Sun Microsystems जैसे बन सकते हैं, चौंकाने वाला है—और आंशिक रूप से सही भी। इन कंपनियों ने कभी पूरे-के-पूरे तकनीकी वर्गों पर राज किया था। DEC मिनीकंप्यूटरों का सम्राट था। Sun ने इंटरनेट की शुरुआती रीढ़ को शक्ति दी। Cray सुपरकंप्यूटिंग का पर्याय था। फिर भी ये सभी अंततः संरचनात्मक बदलावों के सामने झुक गए: माइक्रोप्रोसेसर, वस्तुकरण (commoditization), खुले मानक, और नए आर्थिक मॉडल।

इस रूपक की ताकत इसमें है कि यह व्यवधान (disruption) के खतरे को उजागर करता है। आज LLMs इसलिए हावी नहीं हैं क्योंकि वे परिपूर्ण हैं, बल्कि इसलिए क्योंकि वे सबसे पहले बड़े पैमाने पर पहुंचे।

लेकिन इस तुलना की एक सीमा भी है। LLMs हार्डवेयर नहीं हैं। यदि उनका पतन होगा, तो वह बंद होती फैक्ट्रियों या नीलामी जैसा नहीं दिखेगा। वह कहीं अधिक सूक्ष्म होगा—ओपन-सोर्स होना, फाइन-ट्यूनिंग, डिस्टिलेशन, प्लेटफॉर्म में विलय, या बुनियादी ढांचे में चुपचाप धँस जाना। Sun “मर” गया, लेकिन Java नहीं। DEC गायब हो गया, लेकिन उसके विचार आधुनिक कंप्यूटिंग में जीवित हैं। इसी तरह, भले ही आज के एआई नेता अपना प्रभुत्व खो दें, उनके मॉडल, तकनीकें और डेटा भविष्य की स्थायी आधारशिला बन सकते हैं।

इन कंपनियों से मिलने वाला असली सबक विलुप्ति नहीं, बल्कि केंद्रीयता का खोना है।


ऑपरेटिंग सिस्टम के रूप में LLMs: एआई की इंटरफेस परत

दूसरा रूपक LLMs को ऑपरेटिंग सिस्टम—CP/M, DOS, Windows, Unix—के रूप में देखता है, जो कच्ची कंप्यूटिंग शक्ति और मानव इरादे के बीच की परत बनते हैं। यह तुलना प्रभावशाली है। आज LLMs एप्लिकेशन नहीं रह गए हैं; वे वह इंटरफेस बनते जा रहे हैं जिसके माध्यम से एप्लिकेशन बनाए, पूछे और संचालित किए जाते हैं। APIs, प्लग-इन्स, एजेंट, को-पायलट—ये उसी तरह हैं जैसे कभी ऑपरेटिंग सिस्टम पर एप्लिकेशन इकोसिस्टम विकसित हुआ था।

Unix यहाँ चेतावनी देने वाला उदाहरण है। दशकों पहले बना होने के बावजूद वह आज भी आधुनिक प्रणालियों की नींव है, क्योंकि वह मॉड्यूलर, पोर्टेबल और अवधारणात्मक रूप से स्वच्छ था। यदि LLMs भी उतने ही मौलिक बन जाते हैं, तो वे आलोचकों की अपेक्षा से कहीं अधिक समय तक टिक सकते हैं।

लेकिन यह रूपक भी टूटता है। ऑपरेटिंग सिस्टम नियतात्मक (deterministic) होते हैं। LLMs संभाव्य (probabilistic) हैं। OS कर्नेल भ्रम (hallucination) नहीं करते। उन्हें लगातार पुनः-प्रशिक्षण की आवश्यकता नहीं होती। उन्हें चलाने के लिए शहर-जितने बिजली संयंत्र नहीं चाहिए। इसके अलावा, OS का प्रभुत्व मानकों (जैसे POSIX) से आया, जबकि एआई आज डेटा, कंप्यूट और संरेखण दर्शन के आधार पर खंडित हो रहा है।

यदि LLMs ऑपरेटिंग सिस्टम हैं, तो वे अभी अस्थिर ऑपरेटिंग सिस्टम हैं—जो विकसित होते हुए ही निर्भरता का केंद्र बन चुके हैं।


एआई युग के TCP/IP के रूप में LLMs

सबसे क्रांतिकारी रूपक पूरी कहानी को पलट देता है: क्या होगा यदि LLMs दृश्य इंटरफेस न होकर अदृश्य प्रोटोकॉल परत हों—Windows से अधिक TCP/IP?

TCP/IP ने दुनिया को इसलिए बदला क्योंकि वह अदृश्य हो गया। उपयोगकर्ता कभी “TCP/IP का उपयोग” नहीं करते थे; वे उसके सक्षम किए गए अनुप्रयोगों का उपयोग करते थे। ईमेल, वेब, स्ट्रीमिंग, क्लाउड—सब उसी पर चुपचाप सवार थे।

यह कल्पना असंभव नहीं कि भविष्य में LLMs भी ऐसे ही हों। वे सस्ते, सर्वव्यापी और अंतर्निहित हो जाएँ—तर्क इंजन, एजेंट, सिमुलेशन और मल्टीमॉडल प्रणालियों को शक्ति देते हुए, बिना कभी चैटबॉट के रूप में प्रत्यक्ष दिखे। इस दुनिया में मूल्य ऊपर की परतों में चला जाएगा: ऑर्केस्ट्रेशन, एकीकरण और अनुभव।

समस्या अर्थशास्त्र है। TCP/IP इसलिए सफल हुआ क्योंकि वह खुला, न्यूनतम और सस्ता था। आज के अग्रणी LLMs बंद, विशाल और महंगे हैं। यदि उन्हें अदृश्य अवसंरचना बनना है, तो वस्तुकरण अपरिहार्य है—और वस्तुकरण किसी भी खाई (moat) का शत्रु होता है। ऐसा हुआ तो आज के नेता जीवित रह सकते हैं, लेकिन उनका मूल्यांकन, शक्ति और कथात्मक प्रभुत्व नहीं।


गायब कड़ी: “वर्ल्ड व्यू”

इन सभी रूपकों के नीचे एक गहरा दावा छिपा है: वर्तमान LLMs वास्तव में दुनिया को नहीं समझते।

वे असाधारण पैटर्न पहचानकर्ता हैं, लेकिन कमजोर तर्ककर्ता। वे सुंदर भाषा लिख सकते हैं, लेकिन कभी-कभी प्राथमिक भौतिकी में असफल हो जाते हैं। यह आलोचना हाशिए की नहीं है। Yann LeCun जैसे शोधकर्ता लंबे समय से कहते आए हैं कि ऑटोरेग्रेसिव भाषा मॉडल में आंतरिक “वर्ल्ड मॉडल”—कारणात्मक, भौतिक और सामान्य-बुद्धि आधारित प्रतिनिधित्व—नहीं होता।

अगली बड़ी छलांग शायद बड़ा मॉडल नहीं, बल्कि गहरा मॉडल होगी: ऐसे सिस्टम जो धारणा, स्मृति, कारण-कार्य, सिमुलेशन और योजना को जोड़ें। वीडियो-प्रशिक्षित मॉडल, एम्बॉडिड एआई, न्यूरो-सिंबोलिक हाइब्रिड और एजेंटिक सिस्टम इस बदलाव के शुरुआती संकेत हैं। Cyc जैसे प्रयास पहले विफल हुए, लेकिन आज अंतर पैमाने, कंप्यूट और डेटा विविधता का है।

विडंबना यह है कि स्थापित कंपनियाँ “वर्ल्ड व्यू” बनाने की सबसे अच्छी स्थिति में हैं—और उसे अपनाने की सबसे कम प्रेरणा में।


एआई का इनोवेटर डाइलेमा

यहीं Clayton Christensen का प्रसिद्ध “Innovator’s Dilemma” सामने आता है।

आज के नेता अपने मौजूदा उपयोगकर्ताओं को खुश रखते हुए उस भविष्य पर दांव लगाते हैं जो उनके वर्तमान उत्पादों को अप्रासंगिक बना सकता है। Google को खोज राजस्व बचाते हुए सूचना-पहुंच को फिर से परिभाषित करना है। OpenAI को API स्थिरता और सीमांत नवाचार के बीच संतुलन साधना है। हर क्रमिक सुधार आवश्यक प्रतिमान-परिवर्तन को टालने का जोखिम रखता है।

फिर भी, आज की एआई कंपनियों के पास Nokia से बेहतर रास्ता है: क्रमिक विकास। प्लग-इन्स, एजेंट, मल्टीमॉडैलिटी और टूल-यूज़ उन्हें धीरे-धीरे संक्रमण की अनुमति देते हैं। असली संकट शायद तकनीकी नहीं, बल्कि आर्थिक है—गिरती कीमतों और बढ़ती प्रतिस्पर्धा के बीच घातीय R&D लागत को कैसे वित्तपोषित किया जाए।


अज्ञात की खोज में नकदी जलाना

वित्तीय दबाव वास्तविक है। अग्रणी मॉडल को प्रशिक्षित करने में पहले ही सैकड़ों मिलियन डॉलर खर्च हो रहे हैं। भविष्य की प्रणालियाँ—विशेषकर वे जो वर्ल्ड मॉडल या AGI-स्तरीय तर्क की ओर बढ़ें—कई गुणा अधिक कंप्यूट माँग सकती हैं। Sam Altman खुले तौर पर ट्रिलियन-डॉलर-स्तरीय अवसंरचना की बात कर चुके हैं।

Mixture-of-Experts, क्वांटाइज़ेशन और डिस्टिलेशन जैसी तकनीकें लागत वक्र को मोड़ सकती हैं, लेकिन ब्लैक-स्वान घटनाएँ मंडरा रही हैं: ऊर्जा सीमाएँ, विनियमन, भू-राजनीतिक झटके, या ओपन-सोर्स ब्रेकथ्रू जो रातों-रात मूल्य शक्ति को नष्ट कर दें।

इस संदर्भ में यह कल्पना अविवेकपूर्ण नहीं कि आज की कुछ एआई दिग्गज कंपनियाँ 2030 के दशक की Compaq बन जाएँ—महत्वपूर्ण, यादगार, लेकिन केंद्रीय नहीं।


वस्तुकरण, ओपन-सोर्स और Linux परिदृश्य

यदि स्थापित कंपनियाँ लड़खड़ाती हैं, तो ओपन-सोर्स मॉडल वही भूमिका निभा सकते हैं जो Linux ने Windows के सामने निभाई थी: कम चमकदार, अधिक खंडित, लेकिन अथक रूप से अनुकूल। Llama, Mistral और उनके व्युत्पन्न पहले ही कुछ क्षेत्रों में बंद मॉडलों से बेहतर प्रदर्शन कर रहे हैं।

यह भविष्य अव्यवस्थित होगा। इंटरऑपरेबिलिटी कठिन होगी। मानक पीछे रहेंगे। लेकिन नवाचार तेज होगा। एआई स्टैक स्वयं इंटरनेट जैसा दिखने लगेगा—परतदार, अराजक और ऊपर से नियंत्रित करना असंभव।


अंतिम प्रतिपक्ष: अनुकूलन को कम मत आँकिए

इतिहास एक आशावादी सबक भी देता है। Microsoft ने DOS, Windows, वेब, मोबाइल और क्लाउड—सभी दौरों में जीवित रहकर दिखाया, न कि पूर्ण भविष्यवाणी से, बल्कि निरंतर अनुकूलन से। IBM ने हार्डवेयर छोड़कर सेवाओं को अपनाया। पतन भाग्य नहीं होता।

यदि आज की एआई कंपनियाँ “वर्ल्ड व्यू” प्रणालियों को—अधिग्रहण, साझेदारी या आंतरिक पुनराविष्कार के माध्यम से—सफलतापूर्वक अपनाती हैं, तो वे अपेक्षा से अधिक समय तक प्रभुत्व बनाए रख सकती हैं। लेकिन गुरुत्व-केंद्र बदलेगा। अगले दशक के विजेता संभवतः मॉडल निर्माता नहीं, बल्कि इंटीग्रेटर होंगे—वे कंपनियाँ जो एआई को उपकरणों, प्लेटफार्मों, विश्वास और वितरण के साथ जोड़ सकें।

अंततः, ये रूपक पतन की भविष्यवाणी नहीं, बल्कि विनम्रता की याद दिलाते हैं। तकनीक में कुछ भी हमेशा केंद्र में नहीं रहता। सबसे शक्तिशाली अमूर्तताएँ भी अंततः अवसंरचना, किंवदंती या फुटनोट बन जाती हैं।

एक ही बात निश्चित है: साधारण चैटबॉट्स का युग पहले ही समाप्त हो रहा है। जो आगे आएगा, वह दुनिया के बारे में केवल बात नहीं करेगा—वह उसे मॉडल करेगा।



Saturday, January 03, 2026

Waters: Bigger Deal Than Mars (10)

Saturday, August 04, 2012

Fred Wilson's Blog: A Gift That Keeps Giving


Fred Wilson has an impressive track record as a VC. That is public knowledge. I have lost count of how many tech companies I got really, really excited about only to later learn it was a Fred Wilson portfolio company.

The dots I am trying to connect in this post is to propose the thesis that Fred Wilson's blog has been fundamental to his very impressive work as a VC. As in, the great work he has managed to do over the past decade he could not have done without his blog. He has a method about his comments section. That is where he goes fishing.

That thesis springs forth a few questions.

One, is it a required? As in should all VCs aspire to blog? I think yes. There are a lot of a A caliber VCs out there who don't blog daily (or ever) like Fred Wilson does. But what I am saying is if you are starting out as a VC today blogging daily is one of the things you can not afford to not do.

Two, other than fishing expeditions is blogging good for your mind? This is a huge yes. I think blogging is for everybody regardless of industry, regardless of what stage they are at in their careers. Blogging is working out for the mind. That is primarily why I blog, speaking just for myself. It also helps with meaningful networking. If Fred did not have a blog, and I did not have a blog, I doubt we would have met to date, I doubt we would have known each other.

Three, could this thesis be extended to tech entrepreneurs? That is a question up in the air. Because there is no A grade tech entrepreneur (or B grade) who blogs daily. Mark Cuban blogs weekly or every few weeks, but I think of him as a retired entrepreneur, he is more of an angel, and besides, his blog does not have Disqus.

I'd love to see some top tech entrepreneurs blog daily like Fred Wilson does. Is that possible? Advisable? I'd hope so. I wish some A grade people running for office did the same.

To Fred's credit he is a top blogger regardless of his track record as a VC. He goes toe to toe with people who are full time bloggers with nothing else going on on the side. I mean, I think Fred could earn a living simply through blogging. Talk about Plan B, not that he needs one.

Fred Wilson is my favorite solo blogger for a few different reasons. But there came a time when I did not read his blog for months. I felt like I was too sucked into it. I needed my space, I thought. But then my recent thought has been not that I should stay away, but that top entrepreneurs he is not invested in should also read his blog.

Fred Wilson, Mark Zuckerberg And Mobile

Once Fred paid tribute to Ben Horowitz on the other coast by saying he - Fred - simply writes checks. As in, Ben has been an entrepreneur. But in some ways simply writing checks has advantages. Away from action you can get better at vision and insight. Zuck struggles with mobile. Fred's blog talks often about mobile. I don't know of books delivering similar wisdom.




Fred Wilson's Impossible Inbox
Fred Wilson: A DJ
Meeting Fred Wilson In Person
A Surprising Blog Post From Fred Wilson
A Mind Blowing Party
Enhanced by Zemanta

Sunday, January 15, 2012

Engaging Mark Cuban

mcubanMark Cuban
in reply to @mcuban

@mcuban You just told me the people who added smarts to the phone will have a much tougher time doing it to TV.
Jan 15 via webFavoriteRetweetReply

paramendraParamendra Bhagat
mcubanMark Cuban
in reply to @mcuban

@mcuban Gigabit broadband will take care of bottlenecks you point out http://t.co/39cK5kom Your observations have great short term value.
Jan 16 via webFavoriteRetweetReply

Mark Cuban, Television, And The Internet

English: Mark CubanImage via WikipediaThis was in the late 1990s. Bill Gates was trying hard to marry television to the internet. He called it WebTV. He failed. This was before broadband became mainstream. And still broadband is not there yet. I think gigabit broadband is where TV and the Internet become one.

This was in the late 1990s. Larry Ellison was after something he called the network computer. You would not have much of anything on the desktop. The network would have all the software you would need. Steve Jobs told him the technology just did not exist to support that. The richness possible on the desktop was leaps and bounds ahead of the richness on the browser. Again, this was before broadband, way before HTML5.

5G + HTML5 = Magic

Two titans were not seeing it straight. Positive spin would say they were futurists ahead of their times.

Mark Cuban Replies To My Tweet
Mark Cuban: Contrarian On The TV Business

The conventional wisdom in the industry is that we are almost there. We nailed the phone. Now TV is next. And we are almost there. Even Steve Jobs says so much in his biography. I finally cracked it, he declares.

Not so fast, says Mark Cuban. By personality Mark Cuban is someone you can expect to take a contrarian stand. As he does now. He makes some good points.

Mark Cuban: The TV Business Keeps Getting Stronger!

This is how I summarized his blog post earlier today in another blog post.

(1) TV shows are high quality stuff. Not just anyone can produce them. People like them.
(2) Video is content king. People like consuming content in video format. Much faster broadband might stand a chance but not the broadband we know. The Internet pipes just are not there yet.
(3) Ease of use is supreme. People want to be able to just turn on and watch. No browse and click.

I think all these points are valid. But by the time we hit universal gigabit broadband all three points will have fallen by the wayside.

(1) There's plenty of great quality music on the web. In fact, all the great music is there.
(2) Faster broadband will mainstream video. Video is already big on the web.
(3) People who design smartphones are better positioned than the cable TV people when it comes to simplifying the video consumption experience. I mean, we could get rid of the remote. Voice control, gesture control. There might even be mind reading.

Mark Cuban though makes a solid point that the TV people are not standing still. They are working hard to ease the complexity from another angle.

It is true that for the masses there are times when you just want to sit back and watch.

Mark Cuban Replies To My Tweet

Mark Cuban: Contrarian On The TV Business

Mark CubanImage via WikipediaI love following the VCs I follow in the blogosphere, but I wish my list was more tilted towards entrepreneurs. The problem is the top entrepreneurs don't blog. Mark Cuban is an exception. He does blog. And the guy sure is opinionated.

I think Mark Cuban just told me the people who added smarts to the phone are going to have a much harder time doing the same to TV. I don't think his stand is definitive. But his stand does give me a glimpse into the complexity of the landscape. Mark Cuban of Broadcast.com fame. I remember when they got bought by Yahoo. I was doing some preliminary work on a dot com that went on to do really well, for two years.

Mark Cuban: The TV Business Keeps Getting Stronger!
We had a policy that we never tried to create hits. That we were always going to go wide and create a reason for people to start watching video online. 17 years later. Yep, its been 17 years since we started Broadcast.com (as audionet.com first), Youtube and others are still doing the exact same thing. ...... Good for them ! Except they are making one huge fundamental mistake, they are trying to create hits. They don’t like the idea that beyond a steady stream of 1 hit wonders they haven’t been able to create a sustainable roadmap to content success. In other words, they have no idea how to drive an audience to specific content. Their hits come out of nowhere. ...... viewing for cable networks has skyrocketed and the amount of traditional tv watched has continued to increase. ..... used to be that only movie companies got output deals ..... Today, TV shows are getting output deals and generating lots of revenue across all the different platforms that show TV shows. Its not just syndication,but those online distributors want to make sure they get the best shows and they are committing up front to buy those shows. An output deal. Found money. ...... The TV business isn’t dead. It really isn’t even morphing. Sure people will watch video online. They will watch it on phones. They will download it. But the videos that online distributors pay the most for will be those that have done the best on traditional TV. Which in turn means more money for the production of shows. ...... Online video is to TV today like DVDs were to Movies in the past. A great revenue source that correlated to the movie’s boxoffice. ...... having to hit the internet button on the remote, or even worse, the input button on the remote will not be the path of least resistance for watching tv. Believe it or not, it will be far too much hassle for most people when compared to just turning on and watching TV the old fashioned way. And on top of that, distributors like Dish, Directv, Charter, Comcast, etc are working hard to improve their guide experiences which will be faster and easier than their online counterparts....... last but not least, MOCA, DLNA and good old fashioned wi fi is always going to be a hassle. No one has perfect wi fi at their apartment or house. It always screws up.
(1) TV shows are high quality stuff. Not just anyone can produce them. People like them.
(2) Video is content king. People like consuming content in video format. Much faster broadband might stand a chance but not the broadband we know. The Internet pipes just are not there yet.
(3) Ease of use is supreme. People want to be able to just turn on and watch. No browse and click.

Sunday, October 31, 2010

Event At Hunch: Gender Talk (3)


I got off the train at Union Square because I was running a little early, maybe a lot early. That gave me more walking distance. I take joy in walking. Right by Union Square on the sidewalk I got randomly stopped by an Ethiopian woman with a Masters from New York University who worked for a children's organization. It was a nice talk. I sent her a Facebook email the following day.

Wednesday, September 01, 2010

Mark Cuban Says Put Your Cash Under The Mattress

This has got to be one of the best Mark Cuban blog posts I ever came across. For a risk taking maven to put together such sane advice, we must really be going through some tough times as a people.
Mark Cuban: The Best Investment Advice You Will Ever Get
I Share Mark Cuban's Passion On The FCC Broadband Plan
Free Is The Future: Picking A Fight With Mark Cuban

Here's my summary of his three points.
  1. Credit cards are no good. Pay them off. 
  2. When in doubt, stick with the cash, don't invest.
  3. Spending 15% less is a better return than any on the stock market. 

Wednesday, March 24, 2010

I Share Mark Cuban's Passion On The FCC Broadband Plan

WEST HOLLYWOOD, CA - NOVEMBER 29:  (FILE PHOTO...Image by Getty Images via Daylife
The FCC Needs to Set Its Sights Higher.. Much Higher (Mark Cuban)

The recent FCC broadband plan has been the talk of the town in the tech blogosphere. (Broader Broadband) There seems to be broad agreement in liking what the FCC has come up with. Some key people have come out saying it is not enough. But nobody seems to be saying what I said in one of Fred Wilson's comments sections: The American people need to revolt like they revolted against the British.

Well, here comes along Mark Cuban saying what the FCC is proposing is not entirely enough. And he is saying it with some passion. Yeah, why stop at 100 megabits per second? That might look a lot now, but not long back 5 megabits per second looked like a lot.

A parallel story is Gmail. Gmail storage looked like a lot when it came out. But soon people started running out of space, at least the power users did.

High speed internet to Cuban is less about video and more about Internet 2. Ride on.

Google has its sights on 1 gigabits per second. And although Mark Cuban is on record wanting to upend the Google search business, here he seems to be in agreement with Google's bandwidth goals.

Mark Cuban is worried about applications that might not show up even when speeds go up. I am not. I think it is inevitable that new applications will show up when super high speed is everywhere.

Cuban, passionate plenty, still does not match my talk. Revolt. Free up the spectrum for the people. There Cuban and I seem to have some disagreements. He is more cautious than I'd like.

Free Is The Future: Picking A Fight With Mark Cuban

Mark Cuban: A Quick Thought on the Viacom/Youtube Lawsuit Disclosures
Don’t Waste the Internet on TV – Protect the Future of the Internet
Should the FCC Reclaim Broadcast Spectrum
 

Reblog this post [with Zemanta]

Sunday, November 15, 2009

Web Search, Real Time Search, Social Search


  • Web search, real time search, social search.The social search space is in major flux.
  • Murdoch, Madoff, Mark Cuban. Cuban is missing the point. He is going back to the era of the Yahoo directories way of doing search. The list of the top 1,000,000 sites is in constant flux as it should be.
  • Twitter and Google. This is a dream come true for me. Index all tweets ever. Index all real time tweets. And then help make sense.
  • Facebook updates are very valuable. I am counting links to news articles and videos among the updates. You have your personal repository that you want to dig into at a later date.
  • If Twitter is your entry point to the web which is your entry point to the world, does that put Twitter in some kind of a central position? Sure.
  • The Google Chrome Operating System is showing up earlier than I had expected. But I guess they did say some time in Fall, did they not? A Netbook with a large enough screen with a Chrome OS on it, now that would be something. 
  • Unless Twitter grows and keeps adding new features to richen the experience, it might stay among the tech elite.
  • If this is not JP Rangaswami
  • Give India time. There is space for many Silicon Valleys across the world.
  • A Yahoo Connect sounds like a good idea.



Reblog this post [with Zemanta]

Friday, August 28, 2009

CubanSpeak: Controversial Take On Entrepreneurs


Success & Motivation: What Entrepreneurs Should NOT Do Mark Cuban

If you got a billion dollars, and you claim you have people pitch business plans to you every day, I believe you. But I don't have to take the rest of the rant.

Cuban's blog post tells me it is as hard to be an investor as it is to be an entrepreneur. How do you figure out which horse to bet on? Even the obvious ones are not obvious, or Yahoo would have bought Google when Google offered to be bought not long after launch.

If you want to keep your money safe, keep saying no. You will have your treasure minus the inflation. But if you are wanting to grow your money, or even grow it like crazy, then not being able to find that dark horse must really itch.

And as for tall talk, only a handful of companies engage in paradigm shift products and services in each generation. It is hard to spot them, it is hard to get in once you do spot them. But a young Steve Jobs always talked in terms of changing the world, and changing the course of world history. He did deliver. He invented an industry.

How do you separate the wheat from the chaff? That is harder than picking the winning stocks on the stock market. You need instincts, instincts that deliver.

On Business Models: Free Is Not Always Good
Free Is The Future: Picking A Fight With Mark Cuban

Reblog this post [with Zemanta]