TickerTain
$SPCX····$NVDA····$MSFT····$GOOGL····$ANTHROPIC·$TSLA····$CRWV····$NBIS····$HIMS····$TSM····$ASML····$FIGURE·$LYFT····$UBER····$SNOW····$META····$ELEVENLABS·$CURSOR·$COGNITION·$OPENAI·$AAPL····$RKLB····$XAI·$AMZN····$AMD····$000660.KS····$INTC····
← newsroom
$DISCOVERY-LOOP · Why Google's Top Engineers Just Quit
now playing · $DISCOVERY-LOOP
$DISCOVERY-LOOPbullish· highejaaz
Jeff Dean and Sanjay Ghemawat launch Discovery Loop to automate scientific research via recursive AI loops
Two highest-ranked Google engineers (Level 11, only two ever) building startup to automate AI research first, then apply to grand challenges (medicines, solar, water); recursive s…
$NVDA···bullish· mediumejaaz
Jeff Dean's exit from Google TPUs signals major win for Nvidia GPUs
Jeff Dean explicitly criticized Google TPU architecture as not conducive for training diverse AI models and will shift to Nvidia GPUs, implying Jensen Huang is already engaging hi…
$NVDA···bullish· lowejaaz
Jeff Dean exiting Google TPUs for Nvidia GPUs signals architecture shift
Google's legendary researcher Jeff Dean explicitly said Google TPUs aren't conducive for training diverse AI models and will shift away; implies Nvidia GPUs become default for his…
$GOOGL···neutral· mediumjosh kale
Josh Kale sees Google business as opportunity despite leadership exodus
Google's core business remains spectacular and a good investment opportunity, but leadership quality is a concern after losing star AI researchers like Jeff Dean and Demis Hassabi…
$GOOGL···neutral· mediumjosh kale
Google shares drop 5% ($180B wipe) as DeepMind leadership exits and star researchers leave
Demis Hassabis stepping down as DeepMind CEO and legendary researchers Jeff Dean and Sanjay Ghemawat leaving to start Discovery Loop signals talent drain; stock sold off but core…
$GOOGL···bearish· mediumejaaz
Ejaaz takes bear case on Google after losing all star AI players
Google has lost all its star AI players including Jeff Dean, Sanjay Ghemawat, and sub-CTOs, which will hurt DeepMind's day-to-day execution and AI competitiveness in the near term.
$GOOGL···bearish· mediumejaaz
Bear case: Google has lost all its star AI players and competitive edge
Key technical leaders (Jeff Dean, Sanjay Ghemawat, sub-CTO) have departed for recursive self-improvement startups, leaving Google without the talent that built its AI foundation.
$META···bullish· mediumejaaz
Meta's Muse Spark models show strong cost/performance trajectory vs frontier
Meta is releasing models rapidly (Muse Spark 1.1, 1.2) targeting the Pareto frontier of cost vs intelligence per token, competing with Grok on efficiency rather than raw capabilit…
$META···bullish· mediumejaaz
Meta's Muse Spark models target cost-intelligence frontier; Zuck's compute advantage could close gap
Meta is releasing models (Muse Spark 1.1, 1.2) competitive on cost per token vs frontier models; with massive compute scale and aggressive release cadence, Meta could compete with…
$META···bullish· mediumjosh kale
Meta and xAI will compete on intelligence per token, not raw capability
If Meta and xAI can deliver 80-95% of frontier performance at 1/10 to 1/50 the cost, enterprises will route workloads to cheaper models; routing platforms become critical infrastr…
$META···bullish· mediumjosh kale
Josh Kale agrees Meta will compete on cost-efficiency plane with frontier labs
Meta and xAI will eventually compete with OpenAI/Anthropic on a different plane — delivering 80-95% of capability at 1/10 to 1/50 the cost — making routing platforms that mix mode…
$XAIbullish· mediumejaaz
xAI has massive compute advantage; Grok 4.6 imminent with aggressive release cadence
Elon Musk controls the most compute (with Zuck) and is aggressively expanding; if compute scaling laws hold, xAI will eventually produce frontier-competitive models; Grok 4.6 rele…
$XAIbullish· mediumejaaz
xAI's massive compute scale and release cadence position it to catch frontier labs
Elon Musk's xAI (referred to as SpaceX AI) has the most compute alongside Meta and is aggressively expanding it; with Grok 4.6 releasing soon and rapid model cadence, compute scal…
$AAPL···bearish· lowejaaz
Apple losing AI talent war; sued OpenAI over hiring, China charging premium for memory
Apple's lawsuit against OpenAI appears weak (wrong email address, confused names); Siri AI launch imminent but Apple struggling to secure memory supply as China charges more than…
$AAPL···bearish· mediumejaaz
Apple getting slapped in AI race: lawsuit fails, memory supply squeezed, Siri delayed
Apple's lawsuit against OpenAI backfired (lawyers emailed wrong person), China refuses memory supply at reasonable prices, and Siri AI launch is imminent but Apple appears despera…
$DEEPSEEKneutral· mediumejaaz
DeepSeek V4 Flash achieves 105x cost advantage vs GPT-5 but quality lags in practice
Chinese open-source model matches benchmarks at fraction of cost but feels weaker in day-to-day productivity use; raises question of cost-quality tradeoff for enterprises.
$DEEPSEEKbearish· lowjosh kale
DeepSeek V4 Flash benchmarks 105x cheaper but fails real-world productivity test
Despite costing 105x less than GPT-4.5 and matching benchmarks, DeepSeek V4 Flash feels substantively weaker in day-to-day productivity use, with fuzzy output quality that benchma…
$CONDUITbullish· lowejaaz
Ex-OpenAI researcher joins Conduit building non-invasive BCI wristband for thought-to-text
Brain-computer interface startup using wristband (non-invasive) to translate brain signals to LLM; early stage but represents next computing paradigm after smartphones.