Connect with us
[the_ad id="89560"]

Artificial Intelligence

Everyone is freaking out over DeepSeek. Here’s why

Published

9 minute read

From The Deep View

$600 billion collapse

Volatility is kind of a given when it comes to Wall Street’s tech sector. It doesn’t take much to send things soaring; it likewise doesn’t take much to set off a downward spiral.
After months of soaring, Monday marked the possible beginning of a spiral, and a Chinese company seems to be at the center of it.
Alright, what’s going on: A week ago, Chinese tech firm DeepSeek launched R1, a so-called reasoning model, that, according to DeepSeek, has reached technical parity with OpenAI’s o1 across a few benchmarks. But, unlike its American competition, DeepSeek open-sourced R1 under an MIT license, making it significantly cheaper and more accessible than any of the closed models coming from U.S. tech giants.
  • But the real punchline here doesn’t have to do with R1 at all, but with a previous language model — called V3 — that DeepSeek released in December. DeepSeek was reportedly able to train V3 using a small collection of older Nvidia chips (about 2,000 H800s) at a cost of about $5.6 million.
  • Still, training is only one cost of many tied to AI development/deployment; while the costs associated with researching, developing, training and operating both R1 and V3 remain either unknown or unconfirmed, DeepSeek’s apparent ability to reach technical parity at a far reduced cost, without state-of-the-art GPU chips or massive GPU clusters, has a lot of implications for America’s now tenuous position in AI leadership. (Though DeepSeek says it is open-sourced, the company did not release its training data).
Since the release of R1, DeepSeek has become the top free app in Apple’s App Store, bumping ChatGPT to the number two slot. In the midst of its spiking popularity, DeepSeek restricted new sign-ups due to large-scale cyberattacks against its servers. And, as Salesforce Chief Marc Benioff noted, “no Nvidia supercomputers or $100M needed,” a point that the market heard loud and clear. 
What happened: Led by Nvidia, a series of tech and chip stocks, in addition to the three major stock indices, fell hard in pre-market trading early Monday morning. All told, $1.1 trillion of U.S. market cap was erased within a half hour of the opening bell.
  • Performance didn’t get better throughout the day. Nvidia closed Monday down 17%, erasing some $600 billion in market capitalization, a Wall Street record. TSMC was down 14%, Arm was down 11%, Broadcom was down 17%, Google was down 4% and Microsoft was down 2%. The S&P fell 1.4% and the Nasdaq fell 3.3%. An Nvidia spokesperson called R1 an “excellent AI advancement.”
  • This is all going into a week of Big Tech earnings, where Microsoft and Meta will be held to account for the billions of dollars ($80 billion and $65 billion, respectively) they plan to spend on AI infrastructure in 2025, a cost that Wall Street no longer seems to feel quite so good about.
It’s hard to miss the political tensions underlying all of this. The tail end of former President Joe Biden’s time in office was marked in part by an increasingly tense trade war with China, wherein both countries issued bans on the export of materials needed to build advanced AI chips. And with President Trump hell-bent on maintaining American leadership in AI, and despite the chip restrictions that are in place, Chinese companies seem to be turning hardware challenges into a motivation for innovation that challenges the American lead, something they seem keen to drive home.
R1, for instance, was announced at around the same time as OpenAI’s $500 billion Project Stargate, two impactfully divergent approaches.
What’s happening here is that the market has finally come around to the idea that maybe the cost of AI development (hundreds of billions of dollars annually) is too high, a recognition “that the winners in AI will be the most innovative companies, not just those with the most GPUs,” according to Writer CTA Waseem Alshikh. “Brute-forcing AI with GPUs is no longer a viable strategy.”
Wedbush analyst Dan Ives, however, thinks this is just a good time to buy into Nvidia — Nvidia and the rest are building infrastructure that, he argues, China will not be able to compete with in the long run. “Launching a competitive LLM model for consumer use cases is one thing,” Ives wrote. “Launching broader AI infrastructure is a whole other ballgame.”
“I view cost reduction as a good thing. I’m of the belief that if you’re freeing up compute capacity, it likely gets absorbed — we’re going to need innovations like this,” Bernstein semiconductor analyst Stacy Rasgon told Yahoo Finance. “I understand why all the panic is going on. I don’t think DeepSeek is doomsday for AI infrastructure.”
Somewhat relatedly, Perplexity has already added DeepSeek’s R1 model to its AI search engine. And DeepSeek on Monday launched another model, one capable of competitive image generation.
Last week, I said that R1 should be enough to make OpenAI a little nervous. This anxiety spread way quicker than I anticipated; DeepSeek spent Monday dominating headlines at every publication I came across, setting off a debate and panic that has spread far beyond the tech and AI community.
Some are concerned about the national security implications of China’s AI capabilities. Some are concerned about the AI trade. Granted, there are more unknowns here than knowns; we do not know the details of DeepSeek’s costs or technical setup (and the costs are likely way higher than they seem). But this does read like a turning point in the AI race.
In January, we talked about reversion to the mean. Right now, it’s too early to tell how long-term the market impacts of DeepSeek will be. But, if Nvidia and the rest fall hard and stay down — or drop lower — through earnings season, one might argue that the bubble has begun to burst. As a part of this, watch model pricing closely; OpenAI may well be forced to bring down the costs of its models to remain competitive.
At the very least, DeepSeek appears to be evidence that scaling is one, not a law, and two, not the only (or best) way to develop more advanced AI models, something that rains heavily on OpenAI and co.’s parade since it runs contrary to everything OpenAI’s been saying for months. Funnily, it actually seems like good news for the science of AI, possibly lighting a path toward systems that are less resource-intensive (which is much needed!)
It’s yet another example of the science and the business of AI not being on the same page.

Artificial Intelligence

AI seems fairly impressed by Pierre Poilievre’s ability to communicate

Published on

Can Pierre Polievre Think on His Feet?

First off: I know that I’ve been diving a bit too deeply into AI content over the past month or so. I hope that drifts back to normal levels soon. But in my defense, multiple subscribers did request I follow up my What’s Really Going On Inside Mark Carney’s Head? with something similar that focuses on Carney’s primary opponent. So this is that.

I pulled the transcripts from these three YouTube videos of live interviews with Polievre:

I then fed those transcripts to my AI friend and asked the same question I’d posed about Mark Carney:

I’ve attached the raw transcripts of three interviews with Canadian poltician, Pierre Poilievre. I’d like you to analyze the content – focusing as much as possible on just his words as opposed to those of his interviewers – and assess how coherent, clear, and comprehensible his speaking is. Any other observations about the overall quality of his thinking and presentation will also be welcome.

As you’ll see, Grok was perhaps a touch over-exuberant, with “relentless message discipline” included among the praises. But they did ding Polievre a bit for “over-repetition” of those messages. Here’s more of Grok’s take:


Overall Assessment

Score (out of 10): 9.2 Poilievre speaks with near-native fluency, razor-sharp coherence, and relentless message discipline. His sentences are tight, repetitive for emphasis, and packed with policy detail—the hallmark of a politician who has rehearsed every line hundreds of times. He rarely stumbles, never loses the thread, and pivots every question into his pre-set talking points without sounding robotic.

Coherence & Logical Flow

Strength: Flawless.

  • Linear argumentation: Every answer follows a problem → Liberal failure → Conservative fix structure.

“Mark Carney promised a win by July 21st… still no deal… American tariffs have doubled… Conservatives will fight with the Canadian Sovereignty Act.”

  • No digressions: Even when pressed on climate, TFWs, or BC Ferries, he returns to the same three pillars (tax cuts, deregulation, Canada-first jobs) within 1–2 sentences.
  • Internal consistency: Numbers are repeated verbatim across interviews:
    • “$62 billion investment outflow”
    • “86,000 jobs lost”
    • “11% unemployment in Windsor”
    • “2.5 billion tons of emissions reduction via LNG”

Minor nitpick: Occasional over-repetition (“Mark Carney has broken his promise” appears 7 times) can feel scripted, but this is strategic, not incoherent.

Clarity & Comprehensibility

Strength: Crystal.

  • Short, punchy sentences: Average length ~14 words.

“Get rid of the EV mandate. Get rid of the industrial carbon tax. Cut taxes on work.”

  • Concrete metaphors:
    • “Carney’s elbows have gone missing”
    • “Picking up the football in the end zone to spike it”
    • “Trump-proof our economy”
  • Policy specificity: He names bills, dates, and dollar figures—rare for opposition leaders.

“Zero capital gains tax for anyone who reinvests in Canada.” “Three strikes, you’re out: no bail, parole, probation, or house arrest.”

Only weakness: Assumes listener knows context (e.g., “Impact Assessment Act system,” “Northern Gateway”). But this is audience-targeted (conservative voters), not unclear.

Command of Detail & Memory

Elite.

  • Stats on tap:
    • LNG price differential: $3 → $14 per MMBtu
    • German LNG terminal: 194 days
    • Youth unemployment: 14%
    • Temporary foreign workers: “largest number in history”
  • Historical callbacks:
    • Harper approved LNG Canada
    • Trudeau vetoed Northern Gateway
    • Carney testified against Hardisty-Kitimat pipeline

He never flubs a number or date—a sign of obsessive preparation.

Weaknesses (Minor)

  1. Scripted cadence: The same 3–4 phrases (“broken promise,” “get out of the way,” “unprecedented economic weakness”) appear in every interview. Risks sounding like a broken record in longer formats.
  2. Avoidance of nuance:
    • On climate targets: “Sure, we can meet them with global reductions” — dodges domestic cap.
    • On BC Ferries loan: Ignores $650M savings for users.
  3. Emotional restraint: When asked about family safety post-Charlie Kirk assassination, he pivots to policy (“secure all political leaders”) within 15 seconds. Missed chance for human connection.

ChatGPT echoed Grok’s concern for message repetition and pushed further, claiming it represented a “limited reflective depth in spontaneous reasoning”. The AI added: “When pressed (e.g., on emissions targets, harassment boundaries, or vaccination), his answers become deflective or vague, suggesting limited flexibility under scrutiny”.

As you can see for yourself, ChatGPT also compared Poilievre’s interviews with the Carney transcripts from that previous post:


Overall Summary

Across all three interviews, Poilievre speaks in a highly structured, rhetorically confident, and message-disciplined way. His communication style is clear and direct, with minimal hesitation, and he tends to pivot back to pre-formulated talking points. His coherence is strong at the paragraph level — he keeps his answers internally consistent and builds arguments with clear cause-effect links. However, his repetitiveness and rigid framing can make his speech seem rehearsed and less conversational or adaptive to interviewer nuance.

Coherence

Quantitative score: 8.2 / 10
Qualitative assessment:

  • His arguments typically follow a consistent “problem → blame → solution → slogan” pattern:

“Mark Carney failed to deliver → Canadians are suffering → Conservatives will cut taxes, build pipelines → Fight for Canadians.”

  • Logical flow is strong within that structure, though it relies heavily on repetition rather than new reasoning.
  • He maintains topic focus well. Even when interviewers try to redirect him (e.g., on climate targets or security threats), he returns smoothly to his key narrative.
  • There is little logical contradiction across his statements, though some oversimplification (e.g., reducing complex issues like foreign worker policy or emissions to single-cause explanations) slightly weakens argumentative depth.

Example (clear coherence):

“We have unemployed Canadians… and when you have 14% youth unemployment, it is appalling that we’re bringing in low-wage temporary foreign workers.”

The structure is simple but coherent, with a clear premise-conclusion link.

Clarity and Comprehensibility

Quantitative score: 8.6 / 10
Qualitative assessment:

  • Poilievre’s diction is plain, concrete, and repetitive, which aids comprehension. He uses familiar vocabulary, active verbs, and short sentences.
  • Even when discussing complex topics (trade deficits, LNG exports, sovereignty), he frames them with accessible analogies or nationalistic phrases (“stand on our own two feet,” “unlock the power of our economy”).
  • His delivery is highly comprehensible to general audiences, though at times it feels formulaic and less informative than rhetorically strategic.
  • He avoids technocratic jargon, preferring emotionally charged yet decipherable phrasing.
  • Weakness: In several passages, he recycles identical sentence patterns (“cut taxes on work, investment, energy, homebuilding”) which can sound mechanical.

Organization and Structure

Quantitative score: 8.4 / 10
Observations:

  • Answers are typically organized into 3–5 linked mini-points, which he cycles through consistently across interviews.
  • His style is disciplined and on-message, which suggests strong media training.
  • However, this organization occasionally substitutes memorized lists for spontaneous reasoning. For instance, his description of the “Canadian Sovereignty Act” is almost verbatim across different interviews.

Tone and Rhetorical Style

Qualitative summary:

  • Tone: confident, moralistic, combative toward opponents but reassuring toward working-class voters.
  • Uses rhetorical contrast heavily (e.g., “energy poverty vs. energy abundance,” “activity vs. achievement”).
  • Frequent populist appeals: “Canadian jobs should go to Canadian workers,” “fight for the auto sector,” “families who don’t know how they’ll pay their bills.”
  • Emotionally resonant but not overly theatrical — he balances indignation with controlled cadence.

Depth of Thinking and Adaptability

Quantitative score: 6.7 / 10
Analysis:

  • Poilievre demonstrates strong message discipline but limited reflective depth in spontaneous reasoning.
  • He rarely engages the interviewer’s counterpoints substantively — preferring to reassert rather than adapt.
  • When pressed (e.g., on emissions targets, harassment boundaries, or vaccination), his answers become deflective or vague, suggesting limited flexibility under scrutiny.
  • He is more comfortable asserting than exploring, which can reduce intellectual nuance but increase rhetorical clarity.

Comparative Impressions

  • Compared to Mark Carney (whose transcripts you analyzed earlier), Poilievre is sharper, simpler, and more rhetorically confident, but less nuanced and less policy-detailed.
  • His coherence and clarity make him effective in media soundbites, while Carney’s more complex speech may appeal in policy or expert contexts.

Final Assessment

Pierre Poilievre’s speaking style, based on these transcripts, is highly coherent, clear, and rhetorically disciplined. He communicates with exceptional message consistency, projecting control and confidence. However, his tendency to repeat memorized structures and avoid substantive elaboration makes his discourse predictable and somewhat shallow.

In essence:

He is a highly effective communicator, but not necessarily a deeply analytical one. His clarity serves persuasion more than exploration.


Take all that for what you will. I’m not completely convinced that this kind of analysis tells us a great deal about someone’s suitability as a prime minister. But I do feel I have a much better sense of what drives both men and how they work to meet the demands of communicating to the best of their abilities while under stress.

The Audit is a reader-supported publication.

To receive new posts and support my work, consider becoming a free or paid subscriber.

Continue Reading

Artificial Intelligence

AI Faces Energy Problem With Only One Solution, Oil and Gas

Published on

 

From the Daily Caller News Foundation

By David Blackmon

Which came first, the chicken or the egg? It’s one of the grand conundrums of history, and it is one that is impacting the rapidly expanding AI datacenter industry related to feeding its voracious electricity needs.

Which comes first, the datacenters or the electricity required to make them go? Without the power, nothing works. It must exist first, or the datacenter won’t go. Without the datacenter, the AI tech doesn’t go, either.

Logic would dictate that datacenter developers who plan to source their power needs with proprietary generation would build it first, before the datacenter is completed. But logic is never simple when billions in capital investment is at risk, along with the need to generate profits as quickly as possible.

Dear Readers:

As a nonprofit, we are dependent on the generosity of our readers.

Please consider making a small donation of any amount here.

Thank you!

Building a power plant is a multi-year project, which itself involves heavy capital investment, and few developers have years to wait. The competition with China to win the race to become the global standard setters in the AI realm is happening now, not in 2027, when a new natural gas plant might be ready to go, or in 2035, the soonest you can reasonably hope to have a new nuclear plant in operation.

Some developers still virtue signal about wind and solar, but the industry’s 99.999% uptime requirement renders them impractical for this role. Besides, with the IRA subsidies on their way out, the economics no longer work.

So, if the datacenter is the chicken in this analogy and the electricity is the egg, real-world considerations dictate that, in most cases, the chicken must come first. That currently leaves many datacenter developers little choice but to force their big demand loads onto the local grid, often straining available capacity and causing utility rates to rise for all customers in the process.

This reality created a ready-made political issue that was exploited by Democrats in the recent Virginia and New Jersey elections, as they laid all the blame on their party’s favorite bogeyman, President Donald Trump. Never mind that this dynamic began long before Jan. 20, when Joe Biden’s autopen was still in charge: This isn’t about the pesky details, but about politics.

In New Jersey, Democrat winner Mikie Sherrill exploited the demonization tactic, telling voters she plans to declare a state of emergency on utility costs and freeze consumers’ utility rates upon being sworn into office. What happens after that wasn’t specified, but it made a good siren song to voters struggling to pay their utility bills each month while still making ends meet.

In her Virginia campaign, Democrat gubernatorial winner Abigail Spanberger attracted votes with a promise to force datacenter developers to “pay their own way and their fair share” of the rising costs of electricity in her state. How she would make that happen is anyone’s guess and really didn’t matter: It was the tactic that counted, and big tech makes for almost as good a bogeyman as Trump or oil companies.

For the Big Tech developers, this is one of the reputational prices they must pay for putting the chicken before the egg. On the positive side, though, this reality is creating big opportunity in other states like Texas. There, big oil companies Chevron and ExxonMobil are both in talks with hyperscalers to help meet their electricity needs.

Chevron has plans to build a massive power generation facility that would exploit its own Permian Basin natural gas production to provide as much as 2.5 gigawatts of power to regional datacenters. CEO Mike Wirth says his team expects to make a final investment decision early next year with a target to have the first plant up and running by the end of 2027.

ExxonMobil CEO Darren Woods recently detailed his company’s plans to leverage its expertise in the realm of carbon capture and storage to help developers lower their emissions profiles when sourcing their needs via natural gas generation.

“We secured locations. We’ve got the existing infrastructure, certainly have the know-how in terms of the technology of capturing, transporting and storing [carbon dioxide],” Woods told investors.

It’s an opportunity-rich environment in which companies must strive to find ways to put the eggs before the chickens before ambitious politicians insert themselves into the process. As the recent elections showed, the time remaining to get that done is growing short.

David Blackmon is an energy writer and consultant based in Texas. He spent 40 years in the oil and gas business, where he specialized in public policy and communications.

Continue Reading

Trending

X