TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
←

jeff barr

T3 · host / generalist

Leads AI strategy and model evaluation at Clay, a company building AI-powered go-to-market engineering tools. Oversees internal evals, model selection, and migration to open-weight models for production workloads.

4 calls·4 names·50% bull·last heard 2 months ago·The Information
track record

no scored calls yet — needs a stated position or a categorical verdict, with a matured window vs SPY

top calls

highest conviction · one per company
1sthigh conviction
$ANTHROPICAnthropic

Anthropic Opus 5 matches GPT-5.6 performance at one-third cost per Clay evals

Jeff Barr's internal evaluations at Clay show Opus 5 replicates Anthropic's public benchmarks, delivering near-Fable (GPT-5.6) performance at roughly 33% of the cost, and is replacing Sonnet 4.6 workloads as a spiritual successor.

The Information2026-07episode →
2ndhigh conviction
$ZHIPUZ.ai

GLM 5.2 tops Clay evals at fraction of OpenAI cost, accelerating open-weight adoption

GLM 5.2 sometimes ranks at the top of Clay's internal evaluations while costing a true fraction of even cheaper OpenAI models, making it a serious contender for frontier workloads and driving rapid migration to open-weight models.

The Information2026-07episode →
3rdmedium conviction
$OPENAIOpenAI

OpenAI Fable retains superior writing vibes but faces cost pressure from Opus 5

Fable (GPT-5.6) maintains best-in-class writing quality and 'immaculate vibes' that users prefer for daily work, but Anthropic's Opus 5 and open-weight models are undercutting on price for comparable tier-two workloads.

The Information2026-07episode →

most discussed · click a bar to filter

  • $OPENAI
  • $ANTHROPIC
  • $MOONSHOT-AI
  • $ZHIPU

recurring themes

  • Open Source AI1
  • AI Economics & Business Models1
  • AI Infrastructure1
4 total
$OPENAI
OpenAI
MEDjeff barr·The Information·2 months ago·Claude Opus 5 vs ChatGPT 5.6
OpenAI Fable retains superior writing vibes but faces cost pressure from Opus 5
Fable (GPT-5.6) maintains best-in-class writing quality and 'immaculate vibes' that users prefer for daily work, but Anthropic's Opus 5 and open-weight models are undercutting on price for comparable tier-two workloads.
"I think the vibes on Fable are just so immaculate and just sort of effervescent. Like there's just a there's something to the quality of the writing and the the sort of you know i…"
2:29
$ANTHROPIC
Anthropic
HIGHjeff barr·The Information·2 months ago·Claude Opus 5 vs ChatGPT 5.6
Anthropic Opus 5 matches GPT-5.6 performance at one-third cost per Clay evals
Jeff Barr's internal evaluations at Clay show Opus 5 replicates Anthropic's public benchmarks, delivering near-Fable (GPT-5.6) performance at roughly 33% of the cost, and is replacing Sonnet 4.6 workloads as a spiritual successor.
"the internal benchmarks that we had pretty much matched what Anthropic said publicly, which is really close to Fables performance at a fraction of the cost. And we were seeing I t…"
0:37
$MOONSHOT-AI
Moonshot AI
MEDjeff barr·The Information·2 months ago·Claude Opus 5 vs ChatGPT 5.6
Moonshot KamiKaze 2.6 deployed in Clay's production agents for coding and research
KamiKaze 2.6 is now running in Clay Agent (research) and Sculptor (go-to-market engineering) as part of Clay's expanding open-weight model stack, though GLM 5.2 shows stronger eval performance.
"we use GLM 5.2 and KamiKaze 2.6 pretty extensively at Clay now. Um we've just launched it in Clay agent, which is our main research agent, but we also use it in Sculptor, which is…"
4:32
$ZHIPU
Z.ai
HIGHjeff barr·The Information·2 months ago·Claude Opus 5 vs ChatGPT 5.6
GLM 5.2 tops Clay evals at fraction of OpenAI cost, accelerating open-weight adoption
GLM 5.2 sometimes ranks at the top of Clay's internal evaluations while costing a true fraction of even cheaper OpenAI models, making it a serious contender for frontier workloads and driving rapid migration to open-weight models.
"we use GLM 5.2 and KamiKaze 2.6 pretty extensively at Clay now... for GLM 5.2 sometimes it's at the top of our e-vals, uh but it's truly a fraction of the cost of even, you know,…"
4:32
8
Open Source AItailwind
Clay rapidly migrating frontier workloads to open-weight models like GLM 5.2
Clay's head of AI reports open-weight models have moved from peripheral to serious contenders for frontier workloads within a year, with GLM 5.2 matching or beating proprietary models on evals at a fraction of inference cost, driving material savings for high-volume inference spend.
7
AI Economics & Business Modelsmixed
Anthropic abandoning low-end tier to move upmarket as open weights compress pricing
Anthropic appears to be dropping Haiku 5 and repositioning Opus 5 as the new workhorse tier (replacing Sonnet 4.6), while open-weight models like GLM 5.2 undercut both on cost, forcing a structural repricing of the model stack.
6
AI Infrastructuretailwind
Inference providers Baseten and Fireworks enable secure US hosting of foreign open-weight models
Clay hosts Chinese-origin models (GLM, KamiKaze) via US-based inference providers Baseten and Fireworks, mitigating data sovereignty concerns while capturing open-weight cost advantages — a replicable pattern for enterprise adoption.