Get the Check

If you’ve also been getting distracted by the world cup in the office like the hosts have, you can further distract yourself by also trading the World Cup on Kalshi. Sign up at kalshi.com/sign-up?referral=getthecheck and get $10 when you trade using our code.
This week the pod breaks down Palantir's new AI partnership with Nvidia, the reported plan for OpenAI to hand the US government a 5% stake (for free??), and if Etched's chip is vaporware or has $1B in credible customer pipeline already.
They start with Palantir and Nvidia's plan to build sovereign AI on Nvidia's open Nemotron models, letting customers train and fine tune on top of Palantir while keeping the actual model weights. Most people don’t know this but Nvidia has models. Alex Karp went on a CNBC tour making the case that using closed frontier models costs you more than you think, not in dollars, but in the data you give up. The pod is split on how good that argument actually is (most enterprises already run zero data retention), but they agree Karp's real move was picking a very public fight with OpenAI and Anthropic. Palantir's stock ripped 9% on the news.
Next, the pod gets into OpenAI's reported plan to give the US government a 5% stake, worth roughly $42 billion at its $852 billion valuation. An idea that started as Bernie Sanders (with AOC's backing) floating 50% ownership. The hosts say either Sam Altman got scared of Bernie (probably not) or knows he needs Trump’s favor pre-IPO. They dig into whether this is an effective wealth redistribution tactic, using Alaska's oil dividend fund as the closest comparison. Some how Maya and Anika end up in an argument about open source models by the end of this segment even though that was supposed to be more relevant to the first story haha.
Finally, the Get the Check segment covers Etched, the chip startup founded by Harvard dropouts and Thiel Fellows that just exited stealth after four years, $800 million raised, and over a billion in signed contracts, but nothing actually shipped yet. Maya says it’s all hype and she wouldn’t invest…not that they said she could. Maya keeps coming back to Nvidia's CUDA lock in as the real moat nobody's cracked.
Follow the pod on Instagram and X @getthecheckpod.
00:00 The ideal number of PTO days01:46 Hot take on why people mix Priya and Maya up03:14 The world's most controversial company teams up with the world's biggest company05:03 What the Palantir Nvidia partnership actually does05:44 Alex Karp's controversial CNBC appearance13:34 Why Palantir stock is ripping again after 52 week lows14:59 OpenAI is reportedly giving 5% to the government for free??17:01 The real reason OpenAI is doing this voluntarily17:44 What Trump supporters will think19:35 What Alaska's wealth fund teaches us about public wealth funds22:19 Maya and Anika debate what a good wealth fund would look like26:33 Get the Check segment: Etched34:53 Why CUDA lock in kills every chip startup

In this episode of Get the Check, Maya and Anika kick things off debriefing some international miscommunication chaos before diving into GPT-5.6's Mythos-style rollout, the U.S.-Iran ceasefire that isn't really a ceasefire, and Baseten's massive new round.
They start with OpenAI. GPT-5.6 is out, and it's a genuinely good model, but the real story isn't the model itself, it's the government holding it back. The U.S. government is temporarily restricting access, releasing it to trusted partners first, and OpenAI is on the record saying that can't be a long-term strategy. Meanwhile Z.ai dropped GLM 5.2, and tech leaders like the CEO of Vercel are calling it shockingly good. Eric Schmidt is now saying China is only six months behind the U.S. The girls get into why that timeline matters so much: the U.S. can restrict OpenAI and Anthropic's models all it wants, but China is distilling open-source versions of American models that are just as capable and don't come with any of the safety guardrails.
Next, they get into the U.S.-Iran ceasefire and why it’s putting Trump’s power at risk. On Sunday, Iran's foreign minister declared Iran has sole authority over the Strait of Hormuz, while the U.S. started operating out of the Omani side of the Strait. It's not just the U.S. and Iran who can't agree, Iran's own government and military don't seem to be on the same page either. Trump built a brand on staying out of international wars, and now this one is dragging on, with gas prices up roughly 40% since it started and the midterms creeping closer. It's not looking great for Republicans right now. Kalshi, has midterm election markets live right now, Democrats taking the House and Senate is currently the top probability event. Sign up at kalshi.com/sign-up?referral=getthecheck and get $10 when you trade using our code.
Finally, Baseten closed a $1.5B Series F. Baseten sits on top of the hyperscalers, helping companies running inference manage and optimize GPU usage to bring costs down, which sounds simple but is an extremely hard problem. They first started mooning during the DeepSeek moment. Priya breaks down why she thinks Baseten is the next Stripe, and Maya unpacks the three reasons people actually want custom models: faster, cheaper, or smarter, and which one she thinks matters the least. They also get into the margin pressure Baseten is staring down as hyperscalers move into the space and inference costs keep dropping. Tune in to find out if the pod would give Baseten their check, and stick around for Maya's hot take at the end.
00:00 Maya and Anika are having communication issues
03:21 Explaining body mirroring and other neurodivergent tendencies
05:34 GPT 5.6 quietly rolls out
09:29 GPT 5.6 is cheating and deceiving more than any model before it
18:20 People want locally stored models
19:18 Coinbase goes viral for optimizing AI spend
19:54 Chinese models are mogging
20:44 The risk with China catching up
30:01 Republicans are looking cooked for the midterms
31:52 The SF POV on the midterm elections
35:01 Kalshi’s current midterm odds
37:14 Baseten raises $1.5B
37:26 Baseten product
39:08 Why Baseten is similar to Stripe
40:59 Competition and margins
41:53 Maya’s hot take on Baseten’s valuation
Follow the pod on Instagram and X @getthecheckpod.

In this episode of Get the Check, the hosts break down three big stories: Midjourney's surprise pivot into medical body scanning, Snap's sixth attempt at AR glasses, and Taste Labs, the $18.5M seed-funded startup trying to bring ~taste~ to AI.
They start with Midjourney. The AI image generation company, which is bootstrapped and at $500M ARR with only 163 employees, just announced a full-body imaging system that uses sound waves instead of radiation to produce 3D models of your organs. The underlying hardware comes from Butterfly Network, a public company that put palm-sized ultrasound tech onto a single semiconductor chip. Midjourney's flagship location is going to be a 25K sq ft spa in Union Square with hot tubs, saunas, and cold plunges. Priya and Maya are for sure going. The pod digs into the med spa math: Body Spec's Marina location runs scans every five minutes, eight hours a day, across two machines, pulling $2.5–3M in revenue with three employees and a tiny lease. The pod also dives into the controversies around the launch, including what false positives might mean at scale. Roughly 15 to 30% of healthy people scanned by imaging devices can show abnormalities that trigger follow-up testing and surgeries. Also, it's unclear if Midjourney is trying to position this as equal to an MRI, but Maya says she doesn’t think they are. Maya’s hot take is no one has mapped how human bodies degrade over time and this could change that.
Next, Snap Spectacles. It's generation six of the AR glasses Evan Spiegel has been pushing for over a decade, and investors are not having it. The stock dropped 10% on the news. At $2,195 it’s unclear who the target audience is. The pod dives into why the glasses actually are technically impressive (just not commercially).
Finally, Taste Labs, which you probably saw launch on X this week. The startup raised $18.5M to build what is essentially a design evaluation and data labeling business, hiring people with good taste to annotate AI outputs so models learn aesthetic judgment. They all met the founder at a dinner once and think the hate on X is overblown and too philosophical to be relevant to the clear need for a design eval product. The pod thinks Taste helped train Anthropic's Mythos model (the Fable 5 aesthetic is too similar to the Taste Labs launch screenshots to be a coincidence).
This episode is brought to you by:
Kalshi — Sign up at kalshi.com/sign-up?referral=getthecheck and get $10 when you trade using our code.
Follow the pod on Instagram and X @getthecheckpod.
00:00 Maya forgets her laptop
07:30 Midjourney's medical pivot
11:02 Ultrasound vs. MRI vs. CT explained
14:09 Why some doctors have concerns about the tech
22:18 We might need to open a Med Spa
25:41 The data on human health this could unlock
27:49 A billion people still use Snapchat?
30:02 Advertisers hate Snap
33:30 Investors are over Evan Spiegel, but he's trying to go founder mode
34:14 Snap Spectacles gen 6: everything you need to know
37:33 Is Snap trying to compete with the iPhone
38:42 Why AR glasses make you motion sick (your brain thinks it's poisoned)
41:47 Taste Labs raised $18.5M to end AI slop
48:12 Did they help train Anthropic's Mythos model
54:20 Why does tech suddenly care about taste`

In this episode of Get the Check, the pod breaks down three massive stories: the SpaceX IPO, Anthropic's Fable 5 model getting pulled by the US government hours after launch, and Jeff Bezos' new physical AI company Prometheus raising $12 billion.They start with SpaceX. The company IPO'd on Friday at a $1.7 trillion valuation, popped to $170 a share, and now trades above $2 trillion, making Elon the world's first trillionaire. Robinhood literally crashed from the traffic, and 30% of shares were allocated to retail investors, which is unusually high for an IPO. Maya went deep on SpaceX's S1 filing and breaks down their three actual business lines: Starlink (connectivity at 66% margins), Starships (launch at 30–40% margins and climbing), and AI compute. Maya gives her hot take on SpaceX’s actual valuation. Next, the pod gets into Fable 5, which is Anthropic's long-awaited Mythos-class model. It was released Wednesday, June 10th and got pulled by the US government by Friday. The model was held back for two months after Mythos preview dropped in April due to zero-day exploits that could let attackers breach systems without user interaction. Anthropic released Fable 5 with a classifier system that routes unsafe queries to Opus 4.8 or blocks them entirely. The government disagreed on how effective that was and Secretary Lutnick sent a letter to Dario Amodei asking to pull the model from foreign nationals. It's reportedly Amazon, a huge Anthropic investor, flagged concerns to the government. Anthropic's position is that these exploits exist in all frontier models. Finally, the pod covers Prometheus, Jeff Bezos' new physical AI startup that raised $12 billion at a $41 billion valuation, which is the highest valuation a seed round has ever been raised at. Prometheus is targeting physical world problems: manufacturing, jet engines, drug compounds. His co-CEO is Vic Bajaj, former head of Google's life sciences team. The company is still mostly in stealth. But the bigger conversation is Bezos' take on AI and labor: he thinks two-earner households will become one-earner households as AI handles more work. Anika is not buying it. The pod gets into Jevons paradox, why tech CEOs are not the right people to be making predictions about labor economics, and if this is just narrative building to justify Amazon's 30,000 layoffs over the last year.This episode is brought to you by:1. Kalshi — If you want to trade on what Cursor's acquisition price actually closes at, or how fast SpaceX hits $3 trillion, the market's live. Sign up at kalshi.com/sign-up?referral=getthecheck and get $10 when you trade using our code.2. Known — Use code GETTHECHECK at known.com to skip the waitlist and get a free date.Follow the pod on Instagram and X @getthecheckpod.Chapter titles00:00 SpaceX IPO'd and it's the largest IPO ever07:42 The three business lines behind SpaceX's $2 trillion valuation22:23 Fable 5 drops — and gets pulled by the US government31:35 Amazon reportedly flagged Anthropic's model to the government36:39 Anthropic vs. the government: who's right?43:23 Jeff Bezos raises $12B for Prometheus at a $41B valuation45:45 Bezos thinks AI will create one-earner households

In this episode of Get the Check, the pod breaks down Anthropic's RSI blog, the end of the tokenmaxxing era, whether personal AI assistants are finally close, and Supabase hitting a $10B valuation because of the rise of vibe coding.
They start with Anthropic's recursive self-improvement paper, which says 80% of code merged at Anthropic is written by Claude, and that Mythos makes the right research decision 64% of the time. The hosts get into what RSI actually means, is it AI improving its own weights, or just better scaffolding around the same model? Maya argues we haven't seen AI solve a proof or generate a genuinely new hypothesis that a human didn't think of first. Priya thinks lines of code is a terrible productivity metric, but also the only one people can agree on. The consensus: Anthropic isn't saying anything that new, and the Wall Street Journal's "Anthropic pauses AI" headline was rage bait (at least for Maya..)
Next, the tokenmaxxing era is officially over. Uber blew through its entire 2026 AI budget and it’s only June. Cognition announced it will pay companies up to $10 million if Devin doesn't deliver ROI, which is a bold move that makes Cognition the benchmark-setter for its own product. And app-layer companies are quietly moving to open-source models like Qwen to cut spend by 80-90%, which Maya thinks reverses the minute Anthropic drops Mythos, but Anika has a different take
Then there’s Bernie Sanders' American AI Sovereign Wealth Fund Act, which would require OpenAI and Anthropic to hand 50% of their equity to the federal government. His argument is that these models trained on everyone's data without consent. Somehow both sides, the AI companies and politicians, kind of agree in principle but disagree on the mechanism. Maya's take: if these companies just went public, wouldn't that solve it, but once again Anika has a different take lol.
Finally, the pod digs into the personal AI assistant moment - why Town AI ($55M Series A from a16z) and Poke's Apple Business messaging debut both launched this week. The category has taken so long despite being the "obvious" use case.
Finally, Supabase hitting $10B, 40% of recent YC batches are using it, their Lovable partnership auto-provisions databases for vibe-coders, and Maya explains the Firebase migration issues that Supabase was built to fix.
This episode is brought to you by:
Kalshi — Anthropic's IPO odds are live on Kalshi right now. Sign up at kalshi.com/sign-up?referral=getthecheck and get $10 when you trade using our code.
Known — Use code GETTHECHECK at known.com to skip the waitlist and get a free date.
Follow the pod on Instagram and X @getthecheckpod.
Chapters:
00:00 Maya forgot her pre-planned banter (classic)
00:17 Priya found Maya the perfect bag in Singapore
04:05 Anthropic says AI is building itself, but is that actually true?
07:22 Wait can we define RSI please
14:57 Tokenmaxxing is officially over
15:23 Cognition's $10M AI productivity guarantee
19:18 App-layer companies are quietly switching to Qwen
22:59 The WSJ's Anthropic headline was rage bait for Maya
24:24 Bernie Sanders wants to seize 50% of OpenAI and Sam Altman is down to talk about it
26:05 Maya thinks going public already redistributes AI wealth
31:56 The personal AI assistant is still not here
36:08 What actually changed with Poke's Apple Business launch
39:22 Town AI raises $55M
49:45 Supabase hit a $10B valuation