AI and Software
Open-weight models
open models such as Llama, DeepSeek, Qwen and their impact
- Episodes
- 44
- Share
- 4.2%
- Shows
- 42
- Net sentiment
- +0.37
Down from 5.1% of episodes in the weeks before; net sentiment was +0.18.
Bullish 20Mixed 12Neutral 6Bearish 6
2 weeks
Share of each week's episodes, and net sentiment.
0%2%4%
Sep 21Sep 28
Net sentiment
−10+1
Sep 21Sep 28
The episodes behind it
Why each one was bullish, bearish or in between. Most time on the theme first.
Bullish 20
- Open models like GLM and DeepSeek are closing the gap with frontier labs, enabling unrestricted use cases.Inside The Startup Building Uncensored AI (Abliteration AI) | EP 2345 · This Week in Startups · Oct 3
- Locally runnable open models such as Qwen and Mistral are presented as increasingly capable alternatives.E24 Practical Local LLMs: Hardware, Tools, and Trade-offs · Embedded AI Podcast · Oct 2
- Open models enable customization, local deployment and innovation that closed API models cannot offer.Open models and the future of Physical AI with NVIDIA · Practical AI · Oct 1
- Open-weight models trail the frontier by months and let enterprises own custom, cheaper AI systems.Frontier AI Is a Ferrari. Most Companies Need a Model Y | Turing CEO · Sourcery · Oct 2
- Adapting and fine-tuning open small models for specific tasks offers cost and quality advantages over expensive cloud APIs.Why AI Makes Games Worse—and How to Make Them Better · The Data Exchange with Ben Lorica · Oct 1
- Cheaper open-weight models like GLM and Kimi cut costs substantially while preserving most output quality.Ep 872: AI Cost Control 101: Why Your Chatbot Bill Is Becoming a Board-Level Problem (Start Here Series Vol 31) · Everyday AI Podcast – An AI and ChatGPT Podcast · Sep 29
Show 14 moreShow fewer
- Open models like Qwen and GLM are said to approach frontier performance and can be trained to outperform.Should You Train Your Own Model? · Agents Hour · Sep 28
- Enterprises are shifting workloads to open models to control cost, with the quality gap narrowing.OpenAI, Meta and Microsoft Race to Put AI to Work Inside Your Organization · AI Proving Ground Podcast: Exploring Artificial Intelligence & Enterprise AI with World Wide Technology · Oct 3
- Cheap, capable open-weight models like DeepSeek offer strong value and freedom.Lightning in x402 and Agent Payments | Ben Carman SLP779 · Stephan Livera Podcast · Oct 3
- Open models like Kimi K3 and DeepSeek V4 can earn strong inference margins on current hardware.Ep. 034 - Engrams: How DeepSeek Offloads KV Cache to DRAM and SSD (Core Research) | Jordan Nanos, Cam Quilici, Alec Ibarra, Bryan Shan · SemiAnalysis Weekly · Oct 2
- Open models like Qwen, Kimi and GLM come close to frontier quality and can run on local hardware cheaply.1031: Tokenomics: Why Your Agentic AI Bill Is Exploding (and How to Fix It), with Tyler Cox and Ish Shah · Super Data Science: ML & AI Podcast with Jon Krohn · Sep 29
- Enterprises are increasingly willing to explore open-weight models rather than default to big brands.Beyond the God Model | Alex Atallah & Amjad Masad · The a16z Show · Oct 3
- Open models handle most use cases and can run in secure on-prem environments.The Right Chip for the Right Workload: How Inference Speed Shapes the AI Agent Era · Disambiguation · Sep 30
- Open models run on premises at lower cost and are gaining adoption across major clouds.Kevin Carter says China is now favored to win the AI race outright — and he built two ETFs around that bet · Crypto Coin Show · Sep 28
- A new local inference engine with aggressive quantization makes running large models on consumer Macs more practical.The Stack — October 03, 2026 · The Stack · Oct 3
- Open source is seen as critical to AI development, echoing Linux's triumph over Microsoft's opposition.Julie Fredrickson - The Right to Compute: AI Agents, Nuclear Power & Human Data (Ep. 330) · Infinite Loops · Oct 1
- Moore argues open models are how the industry will keep pace and that calls for slowdown or regulation are unhelpful.Soap Box: HD Moore talks OT security, frontier fearmongering and more · Risky Business · Sep 29
- Post-trained open models deliver higher margins and cost efficiency for his company.20VC: $1BN ARR in 18 Months; The Untold Story of Higgsfield | Spending $4M Per Month on Models | Why Moats in AI are BS | Scaling a Content Team to 150 People with Alex Mashrabov · 20VC · Sep 28
- Ek and Friedberg favor open models for broad innovation and cheaper token costs.Daniel Ek: Life After Spotify, Broken Healthcare Incentives, Catching Disease Early & AI's Potential · All-In Podcast · Sep 28
- Ferrari says cheap self-hosted open-weight models handle most of their AI tasks at very low cost.This EU Firm Made Billions Buying Up Forgotten Tech · Prof G Markets · Oct 4
Bearish 6
- Hosts argue open-source model market share has peaked as enterprises prefer closed frontier models.20VC: Instinct Raises $1B at $10B Valuation | AMD Buys Fei-Fei Li's World Labs for $8.2B | Meta Poaches MongoDB's CEO | Bessemer Raises $5.75B | Oura Pulls IPO & Nubank Eyes $8–12B Monzo Takeover · The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch · Oct 1
- Near-frontier open models built partly via distillation could undercut pricing of closed proprietary models.Trump hails 'morally binding' AI deal · Squawk Box Europe Express · Sep 29
- Speakers expect open-weight models to match Mythos-level capability within months, widening attacker access.How Visa Secures Trillions Using AI Agents & Open-Source Harnesses · AI Security Podcast · Oct 2
- A red team report argues GLM 5.3 offers strong cyber capability without meaningful safeguards.The Stack — September 30, 2026 · The Stack · Sep 30
- Hinton's warning and open-source proliferation are cited as serious safety risks the All-In hosts ignore.Trump and the AI Safety Vibe Shift (Robert Wright & Liron Shapira) · Bloggingheads.tv · Oct 2
- Cheaper open models are eroding pricing power of closed frontier model providers.Meta Muse vs Instinct vs OpenAI Dots: Do AI Agents Have Any Moat? · More or Less · Oct 2
Mixed 12
- Hosts see possible six-to-twelve-month open-source restrictions over cyber risk, with industry interests cutting both ways.AI Execs Sign Joint Commitment, Converting Watts to GDP, the Model Consciousness Debate | Diet TBPN · TBPN · Sep 30
- Open-source labs may distill from Chinese models, and distillation pressures frontier labs' moats.#211 - Is AI Distillation Illegal? · Law of Code · Oct 4
- Most US startups use Chinese open models, raising dependence concerns but also broad benefits from open access.Trump Says AI's Only Guardrail Is a "High IQ President" · Pretrained · Oct 1
- Open models enable private local use but also let bad actors run unguarded frontier-level models.How AI changes cybersecurity's balance of power · The Deep View: Conversations · Sep 30
- Hugging Face and Nvidia push open collaboration, while hosts note self-interest in that advocacy.#37 OpenAI/ Hugging Face Post-Mortem (AI doomerism or Marketing?) · Dev and Doc: AI For Healthcare Podcast · Sep 30
- Chinese open models may carry embedded censorship, yet banning them risks harming US competitiveness.Lawfare Daily: The Trump-Xi Summit and U.S.-China AI Competition · The Lawfare Podcast: Patreon Edition · Oct 2
Show 6 moreShow fewer
- Inference firms are seen as strong bets, but open-source market share may have peaked as enterprises prefer closed labs.20VC: Instinct Raises $1B at $10B Valuation | AMD Buys Fei-Fei Li's World Labs for $8.2B | Meta Poaches MongoDB's CEO | Bessemer Raises $5.75B | Oura Pulls IPO & Nubank Eyes $8–12B Monzo Takeover · 20VC · Oct 1
- Chinese labs release open-weight models to gain attention and market share but forgo direct usage revenue.What America and China Fear Most About AI: A Conversation with Kyle Chan and Helen Toner · The Foreign Affairs Interview · Oct 1
- Open source alternatives were tried but fell well short of Jev's benchmarks, though a local open option is still wanted."The Senior Engineer Death Spiral" - Ep 06 · The Next Token · Sep 28
- Kimi K3 gains Western distribution via AWS, though open-weight monetization and trust concerns remain.Is Polymarket A Money Laundering Machine? | LWIT · Beyond The Build · Sep 28
- Open models and on-premises options are an emerging alternative, though Microsoft's integration remains a strong draw.Rich Gibbons Discusses the Real Cost of AI, Copilot Credits and Microsoft Licensing · Mastering Mondays · Oct 4
- Open models trail closed frontier models, though compression techniques from DeepSeek and others keep improving efficiency.#258 - Opus 5.5, Sol and Luna, Muse, DeepSeek-V4.1-Flash, Xi · Last Week in AI · Oct 2
Neutral 6
- Brief mention of cheaper open-weight models as alternatives to expensive frontier models.Constraints that make developers faster · The Stack Overflow Podcast · Oct 2
- Open models currently less capable, but will soon need the same internal monitoring safeguards.Why AI Agents Cheat | Eric Ho (Goodfire) · The MAD Podcast · Oct 1
- Open model governance and China's open-source stack are mentioned as policy questions.Trump-Xi Meeting, China’s Four-Pronged AI Strategy, Path to US-China AI Cooperation with Kristy Loke · AI Proem Podcast · Sep 28
- Discussion of distillation and Chinese open models and whether they count as IP theft.Inside China’s Gray Market for Anthropic, Cracks Emerge in AI’s Data Center Financing · The Information's TITV · Oct 2
- Notes that open Qwen and Kimi models are fine-tuned bases used by smaller labs.Academia is for Ambition — Alex Zhang, MIT · Latent Space · Oct 2
- Brief mention of open models like Qwen and Kimi being trained on harnesses.Academia is for Ambition — Alex Zhang, MIT · Latent Space: The AI Engineer Podcast · Oct 2
Around it
Companies named
Shows that covered it most
- 20VC2
- The Stack2
- Agents Hour1
- AI Proem Podcast1
- All-In Podcast1
- Beyond The Build1
Travels with
Themes in the same episodes, against chance.
Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.
Briefs like this, for everything you follow.
Ethmos puts research agents on your coverage universe — every debate that touches your names, every executive appearance you'd have missed — and files briefs every morning.
Start free14-day free trial