StarZero All shows Try the podcast clipper →

Mar 23, 2023 · 20m · mad

The 2023 MAD (Machine Learning, Artificial Intelligence & Data) Landscape

Matt Turck · 15m spoken Kevin Zhang · 4m spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

Presented by FirstMark partners Matt Turck and Kevin Zhang, the 2023 MAD (Machine Learning, AI & Data) Landscape presentation analyzes key industry trends, contrasting market consolidation and macroeconomic pressures in data infrastructure against the explosive growth and geopolitical tensions surrounding Generative AI.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 77.6% of the talking time here. How this is scored →

Matt as informed peer 0.6 Guest teaching 0.0 Guest disagreement 0.0 Matt pushing back 0.0
05100:0010:0020:000:45–4:14 · Matt as informed peer 5/10 Landscape Evolution: 2012 vs. 2023 Matt opens the presentation with historical context on the MAD landscape growth from 139 to 1416 companies. Kevin co-presents macro financial data regarding tech drawdowns and venture slowdowns in a completely collaborative tone.4:14–6:56 · Matt as informed peer 0/10 Trend 2: Bundling & Market Consolidation Host monologue by Matt detailing market consolidation trends, ARR valuation challenges, and overcrowding in data quality and MLOps categories. Per instructions for monologue segments, host-side metrics are set to 0.6:56–10:24 · Matt as informed peer 0/10 Trend 3: Modern Data Stack Under Pressure Host monologue by Matt presenting pressures on the Modern Data Stack, architectural alternatives like DuckDB and Trino, and convergence of batch and real-time processing. Host-side metrics scored 0 for monologue format.10:24–13:03 · Matt as informed peer 0/10 Trend 5: Generative AI Goes Mainstream & Multimodal Host monologue by Matt highlighting Generative AI mainstream adoption, ChatGPT bar exam benchmarks, and multimodal tools including a demo of his own Synthesia avatar. Monologue segment scores zeroed out.13:03–15:09 · Matt as informed peer 0/10 Trends 6 & 7: GenAI Boom & Political Economy of AI Host monologue by Matt explaining the political economy of AI, tracing transformer origins from Google to OpenAI and noting the shift of R&D from academia to industry. Monologue segment scores zeroed out.15:09–17:19 · Matt as informed peer 0/10 Trend 8: Opportunities in GenAI SaaS & LLMOps Host monologue by Matt discussing application layer dynamics (e.g. Jasper), workflow defensibility, and emerging infrastructure like LLMOps and vector databases. Monologue segment scores zeroed out.17:19–19:46 · Matt as informed peer 0/10 Trend 9: AI Ethics, Safety & Regulatory Backlash Guest/co-speaker monologue by Kevin discussing AI regulation (EU AI Act), copyright issues, and US-China semiconductor geopolitics including domestic Chinese GPU development. Monologue format yields low interaction scores.19:46–20:53 · Matt as informed peer 0/10 Conclusion and Contact Information Concluding housekeeping segment by Matt providing contact information and closing remarks. Standard low scoring for administrative closing.0:45–4:14 · Guest teaching 0/10 Landscape Evolution: 2012 vs. 2023 Matt opens the presentation with historical context on the MAD landscape growth from 139 to 1416 companies. Kevin co-presents macro financial data regarding tech drawdowns and venture slowdowns in a completely collaborative tone.4:14–6:56 · Guest teaching 0/10 Trend 2: Bundling & Market Consolidation Host monologue by Matt detailing market consolidation trends, ARR valuation challenges, and overcrowding in data quality and MLOps categories. Per instructions for monologue segments, host-side metrics are set to 0.6:56–10:24 · Guest teaching 0/10 Trend 3: Modern Data Stack Under Pressure Host monologue by Matt presenting pressures on the Modern Data Stack, architectural alternatives like DuckDB and Trino, and convergence of batch and real-time processing. Host-side metrics scored 0 for monologue format.10:24–13:03 · Guest teaching 0/10 Trend 5: Generative AI Goes Mainstream & Multimodal Host monologue by Matt highlighting Generative AI mainstream adoption, ChatGPT bar exam benchmarks, and multimodal tools including a demo of his own Synthesia avatar. Monologue segment scores zeroed out.13:03–15:09 · Guest teaching 0/10 Trends 6 & 7: GenAI Boom & Political Economy of AI Host monologue by Matt explaining the political economy of AI, tracing transformer origins from Google to OpenAI and noting the shift of R&D from academia to industry. Monologue segment scores zeroed out.15:09–17:19 · Guest teaching 0/10 Trend 8: Opportunities in GenAI SaaS & LLMOps Host monologue by Matt discussing application layer dynamics (e.g. Jasper), workflow defensibility, and emerging infrastructure like LLMOps and vector databases. Monologue segment scores zeroed out.17:19–19:46 · Guest teaching 0/10 Trend 9: AI Ethics, Safety & Regulatory Backlash Guest/co-speaker monologue by Kevin discussing AI regulation (EU AI Act), copyright issues, and US-China semiconductor geopolitics including domestic Chinese GPU development. Monologue format yields low interaction scores.19:46–20:53 · Guest teaching 0/10 Conclusion and Contact Information Concluding housekeeping segment by Matt providing contact information and closing remarks. Standard low scoring for administrative closing.0:45–4:14 · Guest disagreement 0/10 Landscape Evolution: 2012 vs. 2023 Matt opens the presentation with historical context on the MAD landscape growth from 139 to 1416 companies. Kevin co-presents macro financial data regarding tech drawdowns and venture slowdowns in a completely collaborative tone.4:14–6:56 · Guest disagreement 0/10 Trend 2: Bundling & Market Consolidation Host monologue by Matt detailing market consolidation trends, ARR valuation challenges, and overcrowding in data quality and MLOps categories. Per instructions for monologue segments, host-side metrics are set to 0.6:56–10:24 · Guest disagreement 0/10 Trend 3: Modern Data Stack Under Pressure Host monologue by Matt presenting pressures on the Modern Data Stack, architectural alternatives like DuckDB and Trino, and convergence of batch and real-time processing. Host-side metrics scored 0 for monologue format.10:24–13:03 · Guest disagreement 0/10 Trend 5: Generative AI Goes Mainstream & Multimodal Host monologue by Matt highlighting Generative AI mainstream adoption, ChatGPT bar exam benchmarks, and multimodal tools including a demo of his own Synthesia avatar. Monologue segment scores zeroed out.13:03–15:09 · Guest disagreement 0/10 Trends 6 & 7: GenAI Boom & Political Economy of AI Host monologue by Matt explaining the political economy of AI, tracing transformer origins from Google to OpenAI and noting the shift of R&D from academia to industry. Monologue segment scores zeroed out.15:09–17:19 · Guest disagreement 0/10 Trend 8: Opportunities in GenAI SaaS & LLMOps Host monologue by Matt discussing application layer dynamics (e.g. Jasper), workflow defensibility, and emerging infrastructure like LLMOps and vector databases. Monologue segment scores zeroed out.17:19–19:46 · Guest disagreement 0/10 Trend 9: AI Ethics, Safety & Regulatory Backlash Guest/co-speaker monologue by Kevin discussing AI regulation (EU AI Act), copyright issues, and US-China semiconductor geopolitics including domestic Chinese GPU development. Monologue format yields low interaction scores.19:46–20:53 · Guest disagreement 0/10 Conclusion and Contact Information Concluding housekeeping segment by Matt providing contact information and closing remarks. Standard low scoring for administrative closing.0:45–4:14 · Matt pushing back 0/10 Landscape Evolution: 2012 vs. 2023 Matt opens the presentation with historical context on the MAD landscape growth from 139 to 1416 companies. Kevin co-presents macro financial data regarding tech drawdowns and venture slowdowns in a completely collaborative tone.4:14–6:56 · Matt pushing back 0/10 Trend 2: Bundling & Market Consolidation Host monologue by Matt detailing market consolidation trends, ARR valuation challenges, and overcrowding in data quality and MLOps categories. Per instructions for monologue segments, host-side metrics are set to 0.6:56–10:24 · Matt pushing back 0/10 Trend 3: Modern Data Stack Under Pressure Host monologue by Matt presenting pressures on the Modern Data Stack, architectural alternatives like DuckDB and Trino, and convergence of batch and real-time processing. Host-side metrics scored 0 for monologue format.10:24–13:03 · Matt pushing back 0/10 Trend 5: Generative AI Goes Mainstream & Multimodal Host monologue by Matt highlighting Generative AI mainstream adoption, ChatGPT bar exam benchmarks, and multimodal tools including a demo of his own Synthesia avatar. Monologue segment scores zeroed out.13:03–15:09 · Matt pushing back 0/10 Trends 6 & 7: GenAI Boom & Political Economy of AI Host monologue by Matt explaining the political economy of AI, tracing transformer origins from Google to OpenAI and noting the shift of R&D from academia to industry. Monologue segment scores zeroed out.15:09–17:19 · Matt pushing back 0/10 Trend 8: Opportunities in GenAI SaaS & LLMOps Host monologue by Matt discussing application layer dynamics (e.g. Jasper), workflow defensibility, and emerging infrastructure like LLMOps and vector databases. Monologue segment scores zeroed out.17:19–19:46 · Matt pushing back 0/10 Trend 9: AI Ethics, Safety & Regulatory Backlash Guest/co-speaker monologue by Kevin discussing AI regulation (EU AI Act), copyright issues, and US-China semiconductor geopolitics including domestic Chinese GPU development. Monologue format yields low interaction scores.19:46–20:53 · Matt pushing back 0/10 Conclusion and Contact Information Concluding housekeeping segment by Matt providing contact information and closing remarks. Standard low scoring for administrative closing.

speaking balance: gold is Matt, purple is the guest (3 minute bins)

0:00 · Matt 67% · guest 33%0:00 · Matt 67% · guest 33%3:00 · Matt 58.9% · guest 41.1%3:00 · Matt 58.9% · guest 41.1%6:00 · Matt 100% · guest 0%6:00 · Matt 100% · guest 0%9:00 · Matt 100% · guest 0%9:00 · Matt 100% · guest 0%12:00 · Matt 100% · guest 0%12:00 · Matt 100% · guest 0%15:00 · Matt 77.7% · guest 22.3%15:00 · Matt 77.7% · guest 22.3%18:00 · Matt 34.8% · guest 65.2%18:00 · Matt 34.8% · guest 65.2%
Sharpest disagreement ▶ 2:04 Playful bad-cop framing of market crash

Kevin lightheartedly frames himself as the bad cop presenting severe public market drawdowns and funding contractions, serving as the most contrasting framing in a entirely collaborative presentation.

 Hardest push from Matt ▶ 4:14 Reframing persistent market consolidation assumptions

Matt directly challenges the repeated annual assumption that market consolidation must immediately occur, explaining why VCs and founders resisted selling despite overcrowded categories.

 Biggest teaching moment ▶ 18:40 Detailed update on Chinese GPU developments

Kevin educates the audience on Chinese domestic GPU maker MoreThreads raising capital to counter US semiconductor sanctions despite current driver and performance lags.

 Matt holds his own ▶ 13:40 Technical taxonomy of transformer origins

Matt demonstrates deep domain knowledge by detailing how the foundational transformer architecture originated in research at Google before being commercialized by OpenAI and Microsoft.
the scores for every segment, with the reasoning behind each
ChapterTopicMatt as informed peerGuest teachingGuest disagreementMatt pushing backWhy
Landscape Evolution: 2012 vs. 2023 5000 Matt opens the presentation with historical context on the MAD landscape growth from 139 to 1416 companies. Kevin co-presents macro financial data regarding tech drawdowns and venture slowdowns in a completely collaborative tone.
Trend 2: Bundling & Market Consolidation 0000 Host monologue by Matt detailing market consolidation trends, ARR valuation challenges, and overcrowding in data quality and MLOps categories. Per instructions for monologue segments, host-side metrics are set to 0.
Trend 3: Modern Data Stack Under Pressure 0000 Host monologue by Matt presenting pressures on the Modern Data Stack, architectural alternatives like DuckDB and Trino, and convergence of batch and real-time processing. Host-side metrics scored 0 for monologue format.
Trend 5: Generative AI Goes Mainstream & Multimodal 0000 Host monologue by Matt highlighting Generative AI mainstream adoption, ChatGPT bar exam benchmarks, and multimodal tools including a demo of his own Synthesia avatar. Monologue segment scores zeroed out.
Trends 6 & 7: GenAI Boom & Political Economy of AI 0000 Host monologue by Matt explaining the political economy of AI, tracing transformer origins from Google to OpenAI and noting the shift of R&D from academia to industry. Monologue segment scores zeroed out.
Trend 8: Opportunities in GenAI SaaS & LLMOps 0000 Host monologue by Matt discussing application layer dynamics (e.g. Jasper), workflow defensibility, and emerging infrastructure like LLMOps and vector databases. Monologue segment scores zeroed out.
Trend 9: AI Ethics, Safety & Regulatory Backlash 0000 Guest/co-speaker monologue by Kevin discussing AI regulation (EU AI Act), copyright issues, and US-China semiconductor geopolitics including domestic Chinese GPU development. Monologue format yields low interaction scores.
Conclusion and Contact Information 0000 Concluding housekeeping segment by Matt providing contact information and closing remarks. Standard low scoring for administrative closing.

Statements from this episode (12)

Assertion Supported
Turck: The initial 2012 Big Data Landscape included 139 companies
“And this was the first version back in 2012, a 139 companies, which felt like a lot of companies at the time.”
Matt Turck Mar 23, 2023 ▶ 1:07
Assertion Supported
Turck: The 2023 MAD Landscape features 1,416 companies
“A lot more company, 1416.”
Matt Turck Mar 23, 2023 ▶ 1:21
Assertion Not checkable as stated
Matt Turck: Data infrastructure growth slowed down significantly by 2023
“At a high level, it's a little bit of a tale of two cities, or two worlds, where, you know, the world of data infrastructure had a super exciting last, you know, three, four years, but has slowed down quite a bit.”
Matt Turck Mar 23, 2023 ▶ 1:34
Opinion
Turck: The 2023 AI market is a party like it's 1999
“Meanwhile, in the world of ML and AI, it's a party, like, it's 1999.”
Matt Turck Mar 23, 2023 ▶ 1:51
Prediction Not checkable as stated
Turck: Data startup M&A will mostly be private-on-private mergers
“That's going to be, in my opinion, some tokens is going to be a lot of like private on private kind of like merger to create You know, companies with a broader offering.”
Matt Turck Mar 23, 2023 ▶ 5:57
Opinion
Turck: Data observability sector is overcrowded with too many startups
“In my opinion, data quality, data observability, which is, you know, a wonderful opportunity and very much a needed functionality. I think that's just a little too many companies.”
Matt Turck Mar 23, 2023 ▶ 6:35
Prediction Not checkable as stated
Turck: The MLOps startup category is likely to consolidate
“I think MLOps in general, that's, you know, just in our chart, there's probably, I don't know, 25, 30 companies in MLOps. That's probably going to Consolidate.”
Matt Turck Mar 23, 2023 ▶ 6:45
Assertion Not checkable as stated
Turck: Fully managed data platforms emerge to simplify multi-vendor stacks
“This emergence of fully managed data platforms that quite often abstract away a lot of those vendors. Into just one interface. So you deal with one vendor, one contract, and you don't have to worry about the stitching.”
Matt Turck Mar 23, 2023 ▶ 7:43
Assertion Not checkable as stated
Turck: Cost pressures accelerate architectures using S3, Iceberg, and DuckDB
“And then there's the acceleration of like alternative architectures, you know, storage. Why don't you just dump everything in S three, which is very cheap. And then, you know, you could use iceberg on top to add some structure to it. You can, we talked about D…”
Matt Turck Mar 23, 2023 ▶ 7:56
Assertion Supported
Turck: Google invented the Transformer but is playing catch-up
“So in GPT, which stands for generative pre-trained transformers, the T is a transformer architecture that was actually Developed at Google, and the irony is that Google is playing catch-up to Microsoft and OpenAI right now because OpenAI accelerated the develo…”
Matt Turck Mar 23, 2023 ▶ 13:51
Assertion Supported
Turck: Academia has completely lost the AI research battle to industry
“Academia as of now has just completely lost the battle. In AI research and development, the darker dots, and this is a long time scale at the bottom, show that in the early years of the current wave, a lot of the research came out of academia, but today it's p…”
Matt Turck Mar 23, 2023 ▶ 14:50
Prediction Not checkable as stated
Turck: AI app moats will come from workflow, not the models
“Arguably, the competitive advantage is actually not going to be AI, but the competitive advantage is going to be the workflow And the collaboration that you build on top of the AI to do those things, which I think is not any more or less interesting than your …”
Matt Turck Mar 23, 2023 ▶ 16:12
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 400 conversations transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more