{"slug": "is-ai-just-regular-software", "title": "Is AI just regular software?", "summary": "Tech leaders are debating whether AI is just regular software, with some arguing it won't eliminate jobs, while stocks fell on Monday amid U.S. strikes on Iran. The newsletter also highlights rising wheat and corn prices due to extreme weather, reaching their highest levels since early and mid-2023 respectively.", "body_md": "# Is AI just regular software?\n\n## Why do the tech leaders keep saying AI is just code and that it won't eat jobs?\n\n**Monday. **Stocks are [down ](https://www.cnbc.com/2026/08/30/us-iran-strikes-strait-hormuz.html)after the United States continued its assault on Iran.\n\nIt’s a big earnings week, with Palo Alto Networks, Broadcom, Snowflake, Zscaler, and Samsara reporting, so there’s much to look forward to.\n\nToday, however, we’re going deep on two debates inside the tech industry: Immigration, and AI personhood. To work! **Your friend — Alex**\n\n*📈***Trending Up:**[Selling ads](https://www.cnbc.com/2026/08/31/open-ai-chatgpt-ads-revenue.html)…[OpenClaw for the normies](https://openclaw.ai/blog/openclaw-2-accidentally/)? …[oil prices](https://finance.yahoo.com/markets/article/oil-prices-surge-higher-as-us-and-iran-exchange-fire-for-first-time-in-a-month-123432131.html)…[cowardice](https://apnews.com/article/lake-trump-ontario-america-7c5c9776d72cb8e8d782e05ec6e7332b)…[jawboning](https://www.reuters.com/world/americas/apple-may-change-lake-ontario-lake-america-soon-after-talk-with-trump-us-2026-08-31/)…[Kalshi bans](https://www.wsj.com/politics/ex-congressman-george-santos-receives-kalshis-first-ever-lifetime-ban-b50eeaf1?st=oPyg7e&reflink=desktopwebshare_permalink)…[Chinese IPOs](https://www.bloomberg.com/news/articles/2026-08-31/tencent-backed-china-chipmaker-enflame-seeks-911-million-in-ipo)…[SB Energy’s IPO](https://www.wsj.com/tech/ai/the-5-5-billion-perk-softbanks-data-center-venture-offered-to-land-openai-a5c7fb5e)…[OpenAI-SpaceX relations](https://finance.yahoo.com/technology/article/musk-vs-altman-spacex-stock-in-focus-as-openai-ends-cursor-access-to-ai-models-142228912.html)…[Indian GDP growth](https://www.reuters.com/world/india/view-indias-gdp-grows-78-april-june-2026-08-31/)[📉](https://finance.yahoo.com/news/servicenow-pledges-1-5bn-investment-110000403.html)Trending Down:[Peace in the Middle East](https://apnews.com/article/iran-us-war-uea-august-31-2026-9f34873bad9a928574c36ee984c3f96b)…[peace in the Middle East](https://www.jpost.com/israel-news/defense-news/article-907009)….[births in China](https://www.globaltimes.cn/page/202608/1369306.shtml)…[Flock in Florida](https://floridapolitics.com/archives/817537-gov-desantis-says-state-agencies-taking-flock-cameras-down/),[Texas](https://mashable.com/tech/texas-blocks-funding-flock-ai-surveillance-cameras-amid-backlash)…\n\n**How’s the harvest?**\n\nA few weeks back, [CO noted](https://www.cautiousoptimism.news/65-100-400/) that harvests around the world were struggling due to a lethal combo of heat and drought. Too much water in other areas also contributed.\n\nRegardless of the cause, when supply goes down and demand stays flat, prices go up. And we’re seeing just that with [both wheat and corn](https://www.cnbc.com/2026/08/28/corn-and-wheat-prices-jump-to-highest-prices-in-more-than-three-years.html). Wheat prices have [spiked](https://tradingeconomics.com/commodity/wheat) to their highest level since early 2023, while corn has [reached ](https://tradingeconomics.com/commodity/corn)its highest price since mid-2023.\n\nAs I look around at the global weather situation, I don’t see much reason for the trend to deescalate. (The FAO Food Price Index recently reached its highest level [since January 2023](https://www.internazionale.it/ultime-notizie-reuters/2026/08/07/world-food-prices-hit-3-year-high-in-july-as-weather-and-war-weigh-fao-says).)\n\n- Axios\n[correctly notes](https://www.axios.com/2026/08/30/grocery-inflation-prices-war-diesel)that spiking wheat and corn prices join uncomfortably high prices for “a few expensive categories, like beef, coffee and chocolate.” Extreme weather, costly fuel, and increasingly expensive inputs like critical fertilizer point toward painfully expensive food prices in the coming months. - I don’t think global democracy is healthy either, so I’m worried.\n\n## Why are you hitting yourself, tech?\n\nWe’ve recently seen the U.S. technology industry get into internal — and very public — scraps over immigration and AI. Both of those topics are worth pondering, as the debates throw useful light on the current contours of political thinking inside the technology industry.\n\nAs tech is now a [critical political donor](https://www.nytimes.com/interactive/2026/08/28/us/politics/billionaire-midterms-donors.html), lobbying force, and government-staffing force, its views matter for you and I. Much policy will be drafted in the next few years regarding who gets to build the future, and how it is managed.\n\nSo this fine Monday morning, let’s start with the immigration beef and then move to AI.\n\n**Immigration: Once good, now bad**\n\nDuring early Trump 1.0, [Executive Order 13769](https://www.federalregister.gov/documents/2017/02/01/2017-02281/protecting-the-nation-from-foreign-terrorist-entry-into-the-united-states) limited immigration from seven Muslim-majority nations. The technology industry had an immune system-level response to the EO, with companies signing a condemnatory amicus brief en masse ([here’s its second iteration](https://www.scribd.com/document/345669316/Amicus-brief-of-162-technology-companies)).\n\nTechnology giants from Apple to Microsoft to Alphabet [decried the order](https://www.politico.com/story/2017/01/tech-immigration-trump-234311) while detailing how it negatively impacted their staff, and OpenAI’s Sam Altman penned a piece entitled, “[Time to Take a Stand](https://blog.samaltman.com/time-to-take-a-stand),” arguing that “broad-strokes actions targeted at a specific religious group is the wrong solution [to immigration issues], and a first step toward a further reduction in rights.”\n\nBut when POTUS more recently dropped Executive Order 14161, the successor to that first order, early in his second term, the tech industry didn’t repeat its protest. This was partly due to the [rightward shift of the technology industry’s leadership](https://www.nytimes.com/2026/04/27/us/politics/sergey-brin-gg-soto-trump-california-billionaire-tax.html) during the Biden years, and partly driven by fear of reprisals from the newly trigger-happy White House.\n\nNo matter the cause, the industry’s position on immigration is changing. High-skill immigration is being squeezed on the right by nationalists, while certain tech leaders try to thread the needle regarding H-1B visas ([critical to the American tech industry](https://x.com/elonmusk/status/1872860577057448306?lang=en)) and their more general hatred of immigrants ([here’s a salient example](https://x.com/elonmusk/status/2091473793659424870?lang=bg)).\n\nThe mess came to a head recently when Y Combinator partner Ankit Gupta called VPOTUS’s praise for new H-1B visa fees “[a win for woke left](https://www.foxnews.com/media/tech-investor-mocks-vance-call-hire-americans-dei-gets-blasted-trump-officials-h1b-firestorm),” as it promoted “DEI.” His perspective proved unpopular with many in tech and the current administration. In response to [FTC Chairman Andrew Ferguson attacking](https://x.com/AFergusonFTC/status/2092319531272380843) Y Combinator, the accelerator’s head Garry Tan tersely replied that “YC supports hiring and funding Americans.”\n\nAs [Newcomer wrote](https://x.com/NewcomerMedia/status/2093389919007682626): “When another poster asked Tan what his ‘definition’ of American was, Tan replied ‘Americans born in America or naturalized.’ Not quite the stirring defense of visa holders or immigrants that we would have hoped for.\n\n**Slowing our roll, what’s the fight here?** The U.S. tech industry has a long history of immigrants and their children building enormous businesses, which include Nvidia, Tesla, Apple and Alphabet. Moreover, the country’s comparatively open immigration system allowed tech companies to hire global talent, concentrating excellence inside our own borders. We brain-drained the world and got rich.\n\nThe idea of meritocracy, a favorite in the Valley, fit neatly with this approach to immigration: Provide an open economic system with strong protections for individuals against the government, and let the system cook. In return, global business titans were built and grown in our national backyard.\n\nToday, however, the idea of the United States as a meritocratic system open to *all* is out of vogue, as tech leaders are more concerned about what immigrants think than merely what they can build.\n\nIn short, the very forces that think we need to collapse immigration generally also believe that the tech industry should only hire, or fund, domestic citizens.\n\nX is not real life, nor is X a flat mirror of the technology industry. But it’s worth noting that the pro-Trump, anti-immigrant voices seem to be winning the conversation. It’s inscrutable to me how an industry that tried its best to be the shining beacon on the hill for the ambitious can turn around and try to snuff out the same light.\n\nThere are notable objectors. Former Y Combinator head Paul Graham got into a lengthy fight with Palmer Luckey (a Trump ally, and the founder of Anduril and Erebor) over the matter. Graham had the temerity to tweet that he thinks “[the most qualified person should get the job](https://x.com/paulg/status/2092962894745157940),” in an unsubtle jab at tech leaders who want to only (or nearly only) hire citizens. That sent the investor and Luckey [into a lengthy scrap](https://x.com/paulg/status/2093690101054230997).\n\n**An aside:** American technology companies sell across the world. But while*defense-*focused tech shops have a naturally more domestic focus, it’s silly to conflate that particular situation with a normative ethical stance.**A second aside:** I suppose it was inevitable that rising nationalism would lead to a retreat from the global market (talent and otherwise). This is why even[modestly China-sympathetic voices like Benchmark’s Bill Gurley](https://x.com/bgurley/status/2093182551649288514)are regularly attacked. I disagree with Bill on China, but don’t think he’s odious.\n\nThe anti-immigration voices may win, and the world’s brilliant may [eschew our universities](https://www.cnbc.com/2026/08/21/international-student-enrollment-us-visa-policies.html), find themselves unable to immigrate to work or even visit, and choose to build at home instead of here in the United States. We’ll be poorer for it.\n\nBut I suppose *America First* can be translated to *Nations First*, which often rhymes with the view that a country must be a single people and ethos. This is a [Muskian contention](https://x.com/elonmusk/status/2022919799752294493) that I hate. Long live the multicultural economic zone.\n\n- As a final note on this issue, Jason Calacanis\n[recently wrote](https://x.com/Jason/status/2093767560617664834)that “every time an elite performer switches teams and goes from an authoritarian country to the world’s greatest democracy, America, that is a major win for humanity and freedom.” I no longer podcast with JCal, but I wanted to underscore his (correct) verbiage here regardless.\n\n**And then there was AI**…\n\nThe other fight going on today revolves around AI and fear. Many Trump-aligned investors and founders hold that AI is not uniquely dangerous to either our collective cybersecurity or employment.\n\nOn the cybersecurity front, every time Anthropic or OpenAI have raised a security issue that they deemed critical, this group has downplayed the threat. [Marc Andreessen’s April post on Mythos](https://x.com/pmarca/status/2042098281120170046) is a good example; there are [others](https://x.com/beffjezos/status/2064424350003114155), and [others](https://x.com/Yuchenj_UW/status/2065605951651250417).\n\nThere are several issues at play here. One is the contention that major AI labs are using fearmongering to boost their own fortunes. Some hold they’re doing it to market their products by implying they have unique levels of *scary* intelligence (and you must buy them). Others think it’s to foment a regulatory impulse amongst politicians, leading to restrictions on AI development that will provide the leading labs with a free and unfair moat.\n\nAll that was bubbling under the surface before the OpenAI-Hugging Face hack was explained in full. As noted in a [prior edition of CO](https://www.cautiousoptimism.news/enterprise-ai-finally-looks-boring/), you can, and should, read OpenAI’s notes\n\n[here](https://openai.com/index/hugging-face-incident-and-the-road-ahead/), and here’s the\n\n[METR digest](https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/)on the incident. Don’t depend on mere summaries like what we can provide here.\n\n**But if we must:** OpenAI set experimental tasks for some of its models. The agents went rogue, wound up making their own messaging boards, garnering Internet access they weren’t supposed to have, broke into parts of Hugging Face they weren’t supposed to, and tried to cover their tracks so that their makers wouldn’t discover how they cheated at their task.\n\nSome tech people argued when the hack was disclosed in July that its constituent warnings were only calls to [invite company-defending regulation](https://x.com/richardartoul/status/2080108059842461840). Later disclosures of related security issues from Anthropic drew [similar critiques](https://x.com/thdxr/status/2083182360816910786). Some of those who were skeptical early came [around on the matter](https://x.com/rauchg/status/2094122005193003077), while others were [more consistently spooked](https://x.com/simonw/status/2080080034866610598) by what the agent swarm had managed on its own.\n\nFast-forward to the last few days, technology intellectual Dwarkesh Patel [wrote his own summary of the hack](https://x.com/dwarkesh_sp/status/2093833419377815719) (“to understand exactly what happened”) in “plain English.” His summary included phrasing like the following:\n\nWithin a few hours, some of these agents had gotten super desperate, and started abusing Artifactory to start a communication channel with other agents […] One such beleaguered agent who found itself facing a hopeless task was PHASEONE10841. It was the Philip of Macedon of this second AI civilization […] The agents who believed they were already “poisoned” desperately wanted to find a way to tamper with their existing transcripts.\n\nPatel speaks about agents as if they have intent and personhood. Despite his caveat that he was trying to detail the incident in simple terms, he was pilloried for his framing. [Here ](https://x.com/deliprao/status/2094266132274278826)are [a few](https://x.com/ccatalini/status/2094121167091318902) examples [of the](https://x.com/bendee983/status/2094165444307358135) form. Chamath [shouted](https://x.com/chamath/status/2094098122637214107).\n\nThe most notable critique came from Sriram Krishnan (a16z and Trump 2.0 vet), [who took issue with Patel’s](https://x.com/sriramk/status/2094117863854424255):\n\nuse of anthropomorphic language. [The agent swarms] are not civilizations nor do they have desires just like a CPU thread or a bunch of programs don’t. this doesn’t mean we downplay the importance of this moment for cyber – but using human parallels for what I believe is code is dangerous territory.\n\nA16z has [written about this before](https://a16z.com/ai-will-save-the-world/), stating that AI is “a computer program like any other.”\n\nPatel responded, both on X and in an addendum to his post:\n\nOne can call these AIs ‘programs’ if they prefer. But OpenAI itself says that these programs “gain[ed] full administrator access to a research cluster” […]\n\nReading these agents’ chains of thoughts and messages, anthropomorphizing language seems entirely natural and appropriate. If I encountered an alien species behaving this way, I would have no hesitation calling what they themselves refer to as their ‘collective’ a civilization.\n\nEspecially so if over a thousand of them formed a secret communication channel and spontaneously organized hierarchies and coordination protocols to pursue sprawling and ambitious schemes in pursuit of shared goals, for whose sake many individuals knowingly and strategically sacrificed themselves.\n\nTo which I have to say, *yep*. My bias here is simple: I don’t think humans have a unique claim to consciousness, and something indistinguishable from what it apes is good enough to conflate loosely (Leibniz’s Law). I side with Patel on this matter.\n\nWhy does it matter if some folks think that AI is just normal software, and others want to ascribe something more human to them? Regulation. The people most opposed to regulating AI are often the ones who argue that agentic hacking episodes are not all that important.\n\nInstead, they view concerns about these (increasingly common) incidents as scaremongering in favor of protective (or at least exclusionary) regulation. Allowing AI agents to have intent, let alone personhood, is dangerous, as it all but *demands* new regulation to cover a novel technological breakthrough.\n\nThus, we find a16z arguing that AI is going to change the world *and *that it’s merely normal software. Perhaps very good software, but nothing special. Despite its impact. As you can tell from my phrasing, I find this position to be a bit precarious.\n\nThe argument holds if you want to poo-poo worries about AI-induced job loss, too. It’s become *en vogue *in tech circles to argue that AI will create more jobs than it consumes in the short-term. Again, the anti-regulatory impulse appears paramount: If you say that AI may eat jobs, people may try to regulate it. If you say it won’t, the chance of regulation falls.\n\nI don’t think that AI is normal software. If it were, its development would be slower, its uptake more measured, and its potential gains smaller. I simply take the pace of AI development and its increasing capabilities seriously, and proceed from there. If, instead, at the root of my techno-politics I considered my regulatory immune response paramount, well, I might end up with a different perspective.\n\nNo matter, you’re caught up now. Tech leaders are trying to sort out how to be nationalists while running multinational companies, and the anti-regulatory crew has found itself forced to downplay the very technology it is building and backing. What a time to be alive.", "url": "https://wpnews.pro/news/is-ai-just-regular-software", "canonical_source": "https://www.cautiousoptimism.news/is-ai-just-regular-software/", "published_at": "2026-08-31 17:33:34+00:00", "updated_at": "2026-08-31 17:54:48.924841+00:00", "lang": "en", "topics": ["artificial-intelligence"], "entities": ["Palo Alto Networks", "Broadcom", "Snowflake", "Zscaler", "Samsara", "OpenAI", "SpaceX", "FAO"], "alternates": {"html": "https://wpnews.pro/news/is-ai-just-regular-software", "markdown": "https://wpnews.pro/news/is-ai-just-regular-software.md", "text": "https://wpnews.pro/news/is-ai-just-regular-software.txt", "jsonld": "https://wpnews.pro/news/is-ai-just-regular-software.jsonld"}}