Latest Interviews

OpenAI vs. Deepseek vs. Qwen: Comparing Open Source LLM Architectures

Ankit Gupta

OpenAI, Alibaba Cloud, and DeepSeek have each launched significant open-weight language models featuring distinct Mixture of Experts architectures and advanced long-context capabilities. While OpenAI's GPT-OSS prioritizes inference efficiency on consumer hardware, Alibaba's Qwen 3 introduces flexible dense and sparse variants with dual reasoning modes, and DeepSeek's V3.1 achieves superior memory efficiency through Multi-Head Latent Attention. Despite differing engineering strategies for scaling and alignment, all three families demonstrate comparable performance benchmarks derived from trillions of tokens and complex post-training pipelines.

The Sales Playbook For Founders | Startup School

Tom Blomfield

Early-stage B2B founders often stall in non-revenue design partnerships, so this guide outlines a four-stage framework to accelerate sales from initial discovery to signed Annual Recurring Revenue contracts. The strategy replaces vague, months-long pilots with narrow, paid trials that enforce financial commitment and specific ROI metrics, culminating in recurring contracts featuring automatic renewal clauses. By integrating operational tactics like immediate security certification and high-touch onboarding, founders can bypass traditional procurement bottlenecks and establish scalable revenue streams.

The Ultimate AI Video Stack: Up-to-Date Best Tools to Make Content With AI

Justine Moore

Justine, a partner at A16C and the "Venture Twins" creator, outlines a comprehensive workflow for early-stage AI video production that leverages Google Flow for text-to-video generation and Cling 2.1 for consistent character animation. The presentation details how Hedra facilitates precise lip-syncing with voice cloning capabilities while Higgs Field and Crea manage specialized visual effects and multi-model comparison to refine output quality. By synthesizing these distinct tools into a unified pipeline, creators can currently overcome limitations in audio generation and scene continuity despite the field remaining in its nascent, rapidly evolving stage.

How To Get The Most Out Of Vibe Coding | Startup School

Tom Blomfield

The session establishes "vibe coding" as a disciplined engineering practice, recommending specific tool strategies such as using Replit for novices and dual-workflow setups with Cursor and Windsurf for experienced developers. It outlines a rigorous methodology involving architectural planning, modular Git-based implementation, and strict error handling protocols to manage AI-generated code effectively. The discussion further analyzes current model capabilities, highlighting Claude Sonnet 3.7 for implementation and Gemini for planning, while noting expected performance shifts due to training data biases in frameworks like Ruby on Rails.

Colour visions: a German-election lookahead

Rosie Bloor, Jason Palmer, Tom Nuttall, Moeka Iida, Jeff Goddard, Charlotte Kilpatrick

Germany's upcoming federal election features a fragmented parliament where Chancellor candidate Friedrich Merz leads polling but faces complex coalition hurdles due to the AfD's exclusion and the precarious status of smaller parties like Die Linke and the BSW. Simultaneously, Japan is undergoing a cultural reckoning following the resignation of celebrity Nakai Masahiro for assault, sparking public outrage that exposes deep-rooted legal and societal barriers despite recent reforms raising the age of consent. In London, the traditional pie and mash shops that defined Cockney culture since the Victorian era are vanishing due to rising rents and demographic shifts, forcing surviving establishments to diversify menus with vegan options to adapt to changing customer bases.

Scam Inc 2: Opportunity of a lifetime

Zannie Minton-Beddoes, Sulin Wong

Recruited through deceptive job postings, victims from the Philippines, Uganda, South Africa, and South Asia are smuggled into remote Myanmar compounds where they are held captive and forced to execute "Pig Butchering" cryptocurrency scams. These workers endure 16-hour shifts and brutal physical punishment while impersonating wealthy professionals to defraud White Americans aged 40 to 65 of tens of thousands of dollars. Controlled by Chinese-led hierarchies and increasingly industrialized by satellite-confirmed infrastructure, these self-contained operations utilize deepfake technology and strict racial targeting to sustain a vast, transnational fraud network.

How Scaling Laws Will Determine AI's Future | YC Decoded

Garry Tan

This event analyzes the evolution of AI scaling laws, tracing the transition from the 2020 Chinchilla optimization that prioritized data volume over model size to the current era where OpenAI's O3 model demonstrates that intelligence scales linearly with test-time reasoning compute. The discussion highlights industry concerns regarding pre-training data plateaus while arguing that the scaling revolution is shifting toward reasoning depth and unexplored modalities like robotics and chemical modeling. These insights collectively outline a new trajectory for Artificial General Intelligence that moves beyond simple architectural expansion to leverage extended inference and cross-domain applications.

How To Use AI In Your Startup

Brad, Pete, Gustav, Nicola

The Y Combinator Spring Batch application deadline of February 11th offers selected startups a $500,000 investment and critical network access, with a strong emphasis on relocating to the San Francisco Bay Area to stay synchronized with rapidly evolving market conditions. Founders are advised to integrate large language models into their core product or operations rather than merely pivoting for trendiness, leveraging their physical proximity to industry leaders to identify automation opportunities in sectors like healthcare and legacy software replacement. Case studies such as WAPI illustrate how embedding within the SF community enabled an early pivot to voice AI, contrasting with remote companies that failed to adapt due to an informational lag in understanding current technological capabilities.

How do OpenAI’s o1 and o3 models perform complex reasoning?

OpenAI's O1 model utilizes a "chain of thought" reasoning mechanism to solve complex problems by decomposing them into sequential steps, a strategy originally formalized by Google Brain researchers in 2022. This approach enables the system to dynamically identify errors and adjust its logic rather than relying on simple next-token prediction, significantly improving accuracy in tasks requiring multi-step calculation. By explicitly walking through intermediate stages, such as determining total slices and calculating consumption before finding the remainder, the model mirrors human problem-solving to ensure precise outcomes.

YC Partner and Algolia co-founder Nicolas Dessaigne’s advice for founders starting out.

Nicolas Dessaigne

Founders are urged to immediately launch imperfect products rather than waiting for a perfect idea, prioritizing rapid iteration and direct user engagement to validate market assumptions. This approach rejects over-engineering in favor of a "crappy, quick, and dirty" methodology that secures essential feedback through high-touch interactions with early adopters. By accepting short-term technical debt and imperfections, entrepreneurs can quickly pivot to build the correct product and capitalize on the market's forgiving memory.

How much does a vote matter? | Rob Wiblin

Rob Wiblin

This analysis argues that in competitive US presidential elections, the expected value of an informed vote can exceed $1.7 million due to the massive scale of federal spending and the statistical probability of single votes determining outcomes in swing states. By addressing epistemic risks where uninformed voters rely on heuristics, the text demonstrates that individuals with researched policy positions possess a significant advantage over the average voter, thereby validating the rationality of casting a ballot. Consequently, the author recommends that informed citizens prioritize voting in close contests or directing resources toward high-leverage campaigns rather than relying on passive political engagement.

Starting A Company? The Key Terms You Should Know | Startup School

Dalton Caldwell

This session defines critical startup financial mechanics, including burn rate, profitability thresholds, and the structural differences between angel, seed, and Series A funding rounds. It distinguishes between equity instruments like convertible notes and SAFEs while explaining how metrics such as ARR and Total Addressable Market inform valuation and growth strategy. The discussion concludes by outlining the transition from finding product-market fit to scaling operations, ultimately leading toward an IPO or sustained bootstrapping.

The best consumer companies incorporate both virality and network effect to grow organically.

Metcalfe's Law posits that a network's value scales with the square of its user base, transforming a single-user platform like WhatsApp from worthless to globally vital. This dynamic illustrates how the mathematical relationship between nodes and utility drives exponential growth for individual participants. Consequently, consumer companies achieve organic expansion by strategically combining distinct mechanisms of virality and network effects.

How to Use OCaml's Coercion Operator

camels

This session explores standard OCaml's subtyping and coercion mechanisms, emphasizing how the `:>` operator enables compile-time type checks without runtime overhead or data reallocation. The discussion details value inclusion principles for polymorphic variants, recursive subtyping relations, and the specific constraints imposed by private type abbreviations to enforce invariants. While confirming that list coercions remain free at runtime, the presentation notes upcoming analysis of variant annotations and object types in future sessions.

Stripe Head of Design Katie Dill Reviews Startup Websites

Katie Dill, Aaron Epstein

This session establishes core web design principles prioritizing immediate value clarity and mobile responsiveness while critiquing six specific healthcare and B2B technology sites. Reviewers identified critical usability failures across these platforms, including excessive scrolling on Mito Health, visual clutter on Cygnaz, and confusing brand identity on Taiv, which collectively risk user abandonment. The discussion concludes with actionable recommendations to align visual hierarchy, consistent typography, and clear calls-to-action with user scanning behaviors to boost trust and conversion rates.