I do wish that Emacs was more popular with LLM technologies.
LLMs are powerful at dealing with text. And Emacs is highly extensible and typically text-oriented. Already I see people say that LLMs much Emacs much easier to use (since you can ask an LLM to come up with the elisp for you), but I reckon what Emacs provides ought to be useful the other way.
With Emacs, it's common to see people favour bringing various parts of the system all within Emacs. I'm reminded of that when I've seen how useful it can be to provide API access to logging/code/documentation when asking an LLM to troubleshoot.
With LLMs, there are several uses cases I see which are a natural fit for org-mode, and just imitated in markdown. -- e.g. org mode have TODO items, checklists, tables (including spreadsheet functionality), code blocks, tagging/properties.
> I do wish that Emacs was more popular with LLM technologies.
Let's reword this a bit: "I wish Lisp use was more popular with LLMs...", because using Lisp REPL with AI is such a "life hack", I don't get why more people don't do that. When you give AI a "true Lisp REPL" it works wondrously - it stops randomly guessing and starts empirically understanding the problem and the code - saving time, tokens, your mental energy. And with Emacs' text manipulating machinery on top of that, things can get seriously interesting.
I've started experimenting with claude to run end-to-end tests for an emacs package I'm developing. It's incredible.
The way it works:
1. You start emacs in daemon (server) mode.
2. You prompt claude to instantiate an emacs client and write/run tests.
That's it.
Claude will then "pilot" an emacs instance where you can visibly see it running tests. Since almost everything is a first-class function in emacs, and emacs enables almost complete introspection, claude can debug the code in the execution environment. You can also just look at the piloted emacs instance and prompt the agent on what's wrong in the running application state.
This is much more thorough then just having claude write unit tests because many of the issues you might encounter are visual/gui things - which, again, because emacs allows so much introspection, can be examined by looking at the current application state.
I thought this as well, so I spent a few days working with Claude in emacs. Claude managed to totally break my install by suggesting a bunch of "fixes" that were actually not fixes at all. Very poor performance, just hallucinating things left and right.
I don't need Anthropic to break my emacs, I can do that all by myself.
...and then you woke up and realised that Iran never allowed those inspectors to inspect the actually important facilities and started removing inspectors from the country in 2023 under Biden.
During the time of JCPOA (original, between Iran and the P5+1) inspectors had access to where they wanted to go (sometimes with friction, sure) and were able to place tamper resistant / tamper revealing instrumentation, air filters, and spectrometers - effectively creating a data record that could place a stochastic cap on {enrichment level, volume}.
After Trump ripped up that agreement during his first term, withdrawing from the pact in 2018, that was no longer the case - leading to your linked 2023 statement.
The JCPOA was a huge mistake. Even assuming that it truthfully capped enrichment and prevented the development of an atomic bomb, at the same time it enriched the nation and therefore allowed Iran to finance terrorism and ballistic missiles.
That's naive and simplistic. Your counterparty isn't going to agree to a deal if they get nothing in return. "Stop doing nuclear stuff or else" isn't a deal, it's a demand.
And this fiasco proves that the US can't really do enough militarily if Iran just decides they don't want to abide by a deal/demand. Sure, this war was costly and painful for Iran, but they're coming out of it with better terms than they had before the war.
> but they're coming out of it with better terms than they had before the war.
That is precisely the problem. We have not done enough militarily when we clearly had the upper hand in the first few weeks of the war. We sank their boats, destroyed their planes and anti-aircraft batteries and could freely fly all over the country destroying targets at will. The Trump administration has clearly failed to finish what it started and has left Israel and the GCC countries in a worst position than if it hadn't done anything at all. The "deal" is a pathetic outcome.
We should return to combat operations and finish what we started. This is not about invading the country. It is about neutering the regime so that it can't threaten the region and international waterways. The most important aspect of this is bleeding it dry economically by destroying their ability to export oil. Track and sink the "shadow fleet" anywhere in the world, carpet bomb oil production facilities, and destroy bridges and railways that allow trade by land. Make it clear that any damages caused to neighboring countries will be paid for by Iran using their frozen assets.
Biden was elected _after_ Trump broke the deal so it's unclear why you think that tells us anything about what would have happened had the United States honored the treaty.
The crisis exists because Iran is pursuing nuclear weapons. That's the root issue.
JCPOA was never a permanent solution. Under JCPOA, Iran agreed to temporarily cap enrichment for 15 years until 2030. After which, Iran could enrich freely.
JCPOA had no extension framework, the deal could not be extended without being renegotiated. And Iranian officials refused to agree to permanent enrichment caps, they said the 15-year sunset on enrichment was a non-negotiable.
That's a badly misleading framing of the situation.
The U.S. intelligence community said for more than a year that Iran was weeks away from enriching uranium further to bomb grade, said Daryl Kimball, executive director of the Arms Control Association.
"That is extremely worrisome but that is not weeks away from having a nuclear weapon and not weeks away from having a nuclear weapon that can be loaded on a nuclear missile," Kimball said.
"My understanding from non-governmental sources and the internal assessment of the (intelligence community) is that they believe it would take several months or more to fashion the highly enriched uranium bomb grade into a nuclear device, and one to two years to manufacture a small light nuclear device," Kimball said.
Former Iranian Majles member Ali Motahari said in an April 24, 2022 interview on ISCA News (Iran) that when Iran began developing its nuclear program, the goal was to build a nuclear bomb. He said that there is no need to beat around the bush, and that the bomb would have been used as a "means of intimidation" in accordance with a Quranic verse about striking "fear in the hearts of the enemies of Allah."
"When we began our nuclear activity, our goal was indeed to build a bomb,” former Iranian politician Ali Motahari told ISCA News. “There is no need to beat around the bush,” he said.
When asked if saying this publicly will negatively affect the ongoing JCPOA negotiations, Motahari answered: "Nobody notices what I am saying."
He wasn't a member at the time of those statements nor did he have any involvement in their nuclear program. That's misleading to say the least.
Ali Motahari—former member of Iran’s Parliament—clarified that the interview dates back to May 2022, when he neither held a parliamentary seat nor any official role in nuclear affairs.[0]
So besides a single guy in Iran, are there official delegations that have concluded Iran was developing nuclear weapons? The US has had two National Intelligence Estimates, which consists of all 18 agencies in the intelligence community, conclude that they were not developing nukes.
You're misreading your own sources. The 2007 NIE and Gabbard's 2025 testimony both describe a nuclear weapons program that Iran "suspended in 2003". They confirm a nuclear weapons program existed, the opposite of what you're claiming.
And you want an official source: the IAEA concluded in May 2025 that Iran ran an "undeclared structured nuclear program" until the early 2000s using undeclared material. Then it found Iran in formal non-compliance in June of last year.
"Not building a warhead today" and "never pursued weapons" are different claims, and you swapped one for the other.
I never swapped claims. We're talking about nuclear weapons and why the US started a war with Iran this year due to this supposed "crisis" (in your words) but there's no evidence that a threat is imminent. The NIEs were just one data point against your claim; the burden is still on you to show who else agrees with the assesment that this is a crisis.
> IAEA concluded in May 2025 that Iran ran an "undeclared structured nuclear program" until the early 2000s
"nuclear program" is not the same as a weapons program and the IAEA sounding alarm bells over policy violations is not a conclusion that Iran is/was on a nuclear warpath.
Name the civilian use for Iran's 440 kg of 60% enriched uranium. There isn't one. No power reactor uses it (those run on 3-5%), no research reactor needs it (~20% max), and it sits 98-99% of the way enriched weapons-grade. Iran has enough fissile material for multiple bombs with under two weeks of further enrichment. There is no civilian use for this material. Iran enriched this material, under a mountain, in violation of IAEA requirements, at a site it hid from IAEA investigators until it got caught lying, while stonewalling compliance investigators for years.
Even your own source, the 2025 Annual Threat Assessment report, states: "Iran's enriched uranium stockpile is at its highest levels and is unprecedented for a state without nuclear weapons."
What's your justification for Iran producing this material, if not for a nuclear weapons program?
The 2025 unclassified report does not assess that, but the 2024 one did. Money quote:
Iran has greatly expanded its nuclear program, reduced IAEA monitoring, and undertaken activities that better position it to produce a nuclear device, if it chooses to do so.
Also, that 2025 unclassified report was released before the June 2025 enrichment numbers I shared; these numbers would have influenced the report, and you dodged the question:
What's your justification for Iran producing 440kg of highly enriched 60% material, in secret deep under a mountain in violation of IAEA agreements, if not for a nuclear weapons program?
Because they were under sanctions after trump left the jcpoa. They were not enriching other than for civilian use or working towards a bomb during the agreement and for years after trump left it. Sanctions never ended so might as well work on a bomb then?
For unrelated reasons I personally don't believe Iran was ever pursuing a nuclear weapon, just edging enrichment for leverage.
That said, this:
The US has had two National Intelligence Estimates, which consists of all 18 agencies in the intelligence community, conclude that they were not developing nukes.
has little weight given these were / are the same US intell agencies that were blindsided and caught pants down unaware and in the dark about the 1998 India / Pakistan nuclear test exchange.
The core reason those sunset dates exist is because Iranian officials stated that a sunset on enrichment limits was a non-negotiable. They would not sign a deal without them.
Claiming "agreements are renegotiated and extended" is hypothetical. What incentive does Iran have to agree to enrichment caps post-2030? Why would Iran give up its strongest negotiating card, its nuclear program?
Under JCPOA, Iran capped its nuclear program for 15 years in exchange for:
* Global sanctions relief
* $100-150 billion in frozen assets
* Access to the global oil market
Iran in 2030 under JCPOA already has access to all three. The US already played its best cards to get Iran to agree to JCPOA. The US has little new to offer, other than resumed sanctions.
Under a JCPOA extension, why would items like access to the global oil markets (in addition to sanctions) not be part of the negotiations?
And with JCPOA and its possible continuation, that was a joint agreement among a number of countries - in the current situation, it’s just the U.S./Israel (+ them trying to impose their will on other countries to go along with any carrots/sticks).
> Claiming "agreements are renegotiated and extended" is hypothetical.
What's not hypothetical is that, under the deal, they agreed to not enrich until 2030. What's not hypothetical is that Trump abandoned the deal, with nothing to replace it, allowing them to start enriching in 2017 instead.
And if you're going to claim that renegotiation is hypothetical, then you also have to agree that any other possible future outcome, including one in which Iran develops a functional nuclear weapon, is also hypothetical.
I respect your tastes but, as a human, I like the objectivity in the article. I actually couldn't care less about the backstory and I tend to skip articles that wander around without getting to the point. I prefer to read a book when I want that.
In other words, different people sometimes want different things...
Ok, but brevity is the soul of wit. That's part of my point, this is a long form article with no content. If a human had written it, it would simply have been two to four paragraphs instead! "I made a thing, and it was cool the api stayed the same" rather than this really long winded article saying nothing at all.
You essentially made the libertarian argument without realizing it. According to this line of thinking, we should leave as little as possible in the hands of government exactly because it's either bad already or it will eventually be bad. We should then apply an exceptionally high bar to government responsibilities. These would be things that would be even worse in private hands (police for a simple example).
The fallacy of the libertarian argument is the assumption that it’s possible to have a small distributed government trivially doesn’t hold in presence of bigger adversarial monolithic governments elsewhere.
I made sure to steer clear of Tesla and now SpaceX across my investments. It's not that Musk is an incompetent leader or the companies are bad, but his reality distortion field creates valuations and schemes designed to extract as much money as possible from investors rather than to enrich them.
I'm not a Elon hater because of his politics. There are plenty of rich fucks out there with much more malevolent political views, though the Germans know that salute meant. I started making TikTok's about Elon because he tried building a rocket with agile methodologies more suitable for a smartphone app. On top of that, both starship and cyber truck are vanity projects constrained by his ego to embody his vision of what those things should be. He stopped listening to the smart people he has hired.
He overwork his people and burns them out. If he wasn't such an egotistical twat about his psroduct visions, starship might be flying commercial payloads and cybertruck might not be the worst pick up truck ever made.
So yeah, Elon is an incompetent leader. The companies are bad because you'll burn out if you work there.
I’ve always thought that Elon’s “genius” is that he is able to convince smart/capable people to be very loyal to him. I doubt any of his inner circle would comment about him on HN, but I wonder if my theory is correct?
I once beat it in 27 minutes. I didn't know it was supposed to be hard at the time! I briefly checked speed run records now and that time would have put me in the top 50 "no glitches" ranking!
I'm not that good at games... For some reason PoP leveraged some brain circuitry I have with questionable evolutionary value.
Bionic Commando is that game for me. My best time is just under 16 minutes, but I'm generally the worst at video games of any of my friends; it's one of the few platformers I've beaten at all on the NES.
Like most of the South, where I proudly live, it's a place where the poor and rich live very different lives. It has pretty bad places (just like the UK), but it has areas with great quality of life and is far from "horrible".
Being a Chief Technology Officer is no longer the job. We now need an end-to-end AI orchestrator for the corporation in the Chief Agentic Officer role. Dana Lawson is now obsolete and must be let go.
The brutal reality being its probably most capable in that domain. They'll freely admit they're not "the never the smartest guy in the room" and their decision record proves it
Human C-suites will continue to keep their jobs as long as other humans in similar positions continue to buy into their airy-fairy "strategy" stuff. ;)
The enormous question here is if AI works the way these people say then surely competing companies will emerge that are enormously smaller and more agile, thus eating their lunch completely.
The fact this doesn't appear to be happening suggests something different is in play.
One difference is, when the business hits a tough patch, the bank starts asking questions, and whether or not the bank lets it all continue depends primarily on whether there’s a CEO/CFO who:
- Has history with the bank (trust)
- Is willing to put their neck on the line
AI in the current form has no ability to put its neck on the line.
If a CTO said something in the woods, and no one re-posted it, are they still a CTO or just a crazy person screaming in the woods? By amplifying his message you are justifying his job, position, and standing.
This got me thinking on a new way to improve the news: Instead of reporting on the empty words people say, only report on what they have done, example:
a) Elon Musks says SpaceX is the best thing ever worth the most money ever because he is greatest person ever.
as opposed to:
b) SpaceX, run by Elon Musk, is currently Making $X per quarter, they do this by ......
I actually think almost everything she’s describing is basically writing code. I remember Mitchell Hashimoto wrote this blog post saying he wished people didn’t take “infrastructure as code” literally. That “infrastructure as code” really meant something more like “infrastructure as documented specification”
In the agile spirit of “working software over comprehensive documentation” I do think the future of software engineering is writing code. Speaking through code is not a philosophy that’s going away any time soon imo. If anything, code is that much more important than it was before.
Wow, you got agitated. Writing code was never a job. Software engineering is a job where part of the responsibility is to create code. “Writing code” in itself means nothing.
reply