Sh🅰️kespeare @ShakespeareCap

i’m sorry this guy is such a cornball

Carnage @0xCarnagee

a day ago

Anthropic's CEO Dario Amodei: "some of the early companies we gave Mythos to said – this is a super weapon. please don't release this" on leaving OpenAI: "when you feel that you can't trust someone, when you feel that their values are not what they say they are – that makes it

46 140 2K 672K 3K

@usppdd @DP3188144159988 I can never see $AUR ever transitioning to Ouster LIDAR personally. Very very low chance, near 0%.

@crux_capital_ Might never happen but i'm okay to let it go for better r/r elsewhere

@crux_capital_ around the 50 EMA, low ~80s

Genuinely think this is almost free money. NFA

Sh🅰️kespeare @ShakespeareCap

13 hours ago

@crux_capital_ 14-15$ NOK LEAPs is very tempting right now.

1 0 11 2K 4

@crux_capital_ 14-15$ NOK LEAPs is very tempting right now.

I think there’s a huge opening right now for a major U.S. firm to release a truly competitive open-source AI model. DeepSeek has already shown there’s massive demand for powerful, low-cost models, generating around $500,000 a day from API usage alone. But for a lot of institutions in the U.S., the fact that DeepSeek is tied to China creates trust concerns ofc. That leaves a pretty obvious gap in the market. If a trusted U.S. company released a strong open-source model, it could attract a huge developer base while still making serious money. Would completely dwarf the need to have an ugly token heavy model like claude. It would also be good for the broader AI ecosystem of course.

Sold $ MRLN. I still like the company and can see the long-term path to becoming a multi-bagger, but I think the risk/reward has become less compelling relative to other opportunities created by this drawdown. In my view, Merlin is still a 2–3 year story, with a lot that needs to go right and several execution, regulatory, and commercialization milestones still ahead. The upside case is real, but the timing and probability are far from certain. I also think there is a meaningful amount of hype currently embedded in the stock. My experience with companies at this stage is that they often enter a lengthy “prove it” phase, where the market waits for tangible progress before rewarding the equity again. That can mean months, or even a year-plus, of consolidation at lower levels. $AUR and $JOBY are good examples of this in practice. I still think if you want to get upside off autonomous flight, you still have opportunity through $OUST as a picks and shovels for Shield AI's Hivemind, which i think is much higher r/r as OUST has it's foot in lots of doors across defense/space/robotics/etc.

@accounting_ds isn't $OUST the picks and shovels play for this?

I disagree. Aurora’s Driver was built to be vehicle-agnostic: freight now, passenger later. They’re starting with trucking because highways are easier to monetize first, but the tech isn’t limited to trucks. And honestly, on freeways I’d argue Aurora might be better positioned than Waymo. Shown by Waymo cancelling freeway rides to go back into r&d, meanwhile AUR has over 300k miles. Also i would argue it wont be too late for 10+ years, plenty of room for multiple winners in passenger. Passenger AV demand will be huge, hardware costs should keep falling, and more supply should eventually mean cheaper rides.

Observer-less operations are starting within weeks. Big catalyst for $AUR, as they are finally moving from a 'science expirement' to acting on their partnerships with Volvo, Hirschbach, etc. I might sound crazy now, but i genuinely believe this company can be as big as Waymo in the next 10 years. I want you to think, $AUR has always said that they wanted to move into passenger fleet/robotaxis after they dominate trucking. With capability over all edge cases on freeways and cross-country travel, what will stop $AUR after monetizing their giant backlog from launching passenger fleets for cross-city trips (ie. Dallas to Austin)? Even if Waymo can get over their problems, there is more than enough room for multiple robotaxi players. Don't forget, Uber is a long-time investor of $AUR. $AUR can easily be integrated. I think if you trust management's execution for the next 10 years, TAM is limitless.

Sh🅰️kespeare @ShakespeareCap

7 days ago

$AUR. A company with over 300k miles on freeways is worth 14B MC, but Waymo is worth 126B and is still in R&D for freeways? Remember, Waymo’s early father is $AUR’s founder/CEO. He left Google because bureaucracy was slowing things down. Is it me or do you see a disconnect?

0 0 5 2K 1

@Fraudgers12 @hrdktwt Rider

Really great read from Morningstar’s Seth Goldstein on $AUR today. The part that stood out: Autonomous trucking has logged roughly 500,000 miles so far. Morningstar forecasts that number could reach 80 billion miles annually by 2040. That’s a 160,000x increase from where we are today. What makes Aurora interesting is how few serious competitors remain. Over the past several years, many exited the race, went bankrupt, or shifted focus. This is a TOUGH industry to enter. Today, Aurora appears to be one of the very few companies with a credible path to scaling driverless trucking. Most investors say they wish they had the patience to hold companies like Nvidia through the early years. Opportunities like that always look uncertain in real time, especially when the stock stays stagnant. When I look for investments, I look for businesses with a massive TAM, strong technological advantages, and limited competition. $AUR checks every box. Not financial advice, just my opinions on what I believe are great long term investments for the next decade. -T.E.E.

keyboard_arrow_left Previous keyboard_arrow_right Next

$HLIT Jensen Huang just reframed the broadband thesis: “Telecom networks are becoming national AI infrastructure. They connect people, companies, devices and machines — and now they can become the backbone of new AI clouds.” That quote came as NVIDIA announced a gigawatt-scale AI Cloud partnership with SK Telecom in South Korea. Who has already been deeply integrated into SK Telecom / SK Broadband’s broadband infrastructure? Harmonic. Harmonic helped SK Broadband deploy the world’s largest cable IPTV edgeQAM architecture, bypassing traditional CMTS bottlenecks and moving heavy IP video traffic directly from the main network headend to edge devices. Follow the NVIDIA trail. Comcast is working with NVIDIA to turn its low-latency broadband footprint into an AI grid for real-time edge AI inference. Who has already been deeply integrated into Comcast's broadband infrastructure buildout? Comcast’s next-gen broadband rollout is running through $HLIT/Harmonic, with DOCSIS 4.0 expanding to millions of homes and Harmonic’s cOS platform described as an integral part of Comcast’s multi-gig symmetrical broadband strategy. The mapping is clear: Signal 1: NVIDIA is saying telecom becomes AI infrastructure. Signal 2: Comcast is pushing AI to the network edge through Harmonic cOS software and nodes. Signal 3: SK Telecom is building AI clouds on top of Harmonic's edge network architecture installed years ago. Signal 4: Harmonic’s cOS is now deployed with 150 customers, with CUJO AI software enabling operators to activate L4S — a low-latency network technology that gives real-time apps (Think AR/VR, remote control, robotics, real-time inference, and edge AI workloads) priority treatment without crushing total network throughput. Harmonic is already integrated into 150 operators, ~45M nodes deployed, enabling capability of AI edge infrastructure, and still sitting at a 1.5B MC. Bullish on $HLIT.

keyboard_arrow_left Previous keyboard_arrow_right Next

If you use any of these models (ie. Claude, Codex, etc), especially any of the current Opus models or Fable 5, for decently rigorous tasks, you'd realize that these models are STARVING for compute, and how important the power/compute trade really is. I think from that standpoint, people calling pure-play infra players like $NBIS overvalued, are drastically underestimating the amount of power and infastructure we need to create in the next 5+ years in order to see AI truly live up to the hype.

if someone wants to know why exactly i know it's them i could map out the chain of logic

"Comcast’s network is designed to put more computing power physically closer to customers, creating one of the largest and most capable platforms in the U.S. for delivering real-time AI inference with significantly reduced latency, power consumption, and cost. With advanced DOCSIS 4.0 FDX nodes, smart amplifiers, and intelligent gateways across its footprint, Comcast can support real-time AI inference at scale – something traditional centralized, fiber-only, or wireless networks cannot match." Hmm... what company publicly partnered with Comcast engineered those advanced DOCSIS 4.0 FDX nodes that are allowing them to support 'real time inference at scale'? $HLIT

keyboard_arrow_left Previous keyboard_arrow_right Next