After reading today's Wall Street Journal and some other publications about the impact AI is having on white collar jobs, I have a question:

Is there anything that a "procurement professional" (NCMA's beloved term) do that cannot be done faster and as well or better by AI?

I've been experimenting, and the news is not good.

The news is grim.

How to live long and prosper. Any ideas?

This is an AI-generated executive summary of effects on Contracting Officers, Procurement Analysts, and acquisition offices in general related to article titled "How Adaptable Are American Workers to AI-Induced Job Displacement?". Citation is Sam J. Manning and Tomás Aguirre, "How Adaptable Are American Workers to AI-Induced Job Displacement?," NBER Working Paper 34705 (2026). I could not attach the article as it exceeds MB limit, but the entire article is available at https://www.nber.org/papers/w34705. The summary was created since article runs about 54 pages, and that is a lot of material.

This is only one opinion generated from AI of course.

Note: The link has been removed because it does not work.

Executive Summary on AI Displacement.pdf

5 hours ago, Vern Edwards said:

Is there anything that a "procurement professional" (NCMA's beloved term) do that cannot be done faster and as well or better by AI?

Lots of things for the near term such as strategizing with program offices to define and articulate “need,” develop acquisition strategy in collaboration with program offices and stakeholders, conduct negotiations with industry including consideration of tradeoffs, and making source selection decisions after weighing all the evaluation input.

But this is only over the short period. I would like to say AI wouldn’t have the capability to do all that until considerable time has lapsed such as five or perhaps ten years. But that is wishful thinking noting how much AI has progressed in the past single year.

I think the picture is not bleak for those that get in sync with AI and its progression. The acquisition field needs both AI and knowledgeable and flexible personnel (willing to progress and learn as AI develops). Lots of new opportunities will come out that we haven’t envisioned yet which happens with every new technological implemtation.

Here is article I mentioned yesterday. Sorry link did not work.

w34705 Article on AI Displacement.pdf

I saw an episode of 60 minutes a couple of years ago on the AI revolution taking place in China. It showed how deep learning algorithms are used by apps to pull massive data points, allowing everyday users to apply for and get approved for financial loans on their smartphones in less than 10 minutes without ever needing to step foot inside a traditional bank. Kai-Fu Lee describes how this process uses brute-force data and deep learning to instantly analyze risk and automate white-collar banking tasks.

Does anyone think this type of automation is not going to happen in Government contracting?

On 5/20/2026 at 8:34 AM, Vern Edwards said:

Any ideas?

9 hours ago, Don Mansfield said:

Does anyone think this type of automation is not going to happen in Government contracting?

It is happening.

For me the "idea" is a question.- Where is the magic line for actual human interaction that makes for success? Or as borrowed from another thread will the "experience" be accepted and a new standard of success be set?

I use AI but for me I still embrace the social interaction. AI is a supplement not a replacement.

To Don's point at some point sometimes a real loan officer or a real contraring officer has made a difference in a go no go situation at least by my experience. It will remain so I hope but the need for human interaction will be diminished no doudt.

19 hours ago, Don Mansfield said:

Does anyone think this type of automation is not going to happen in Government contracting?

9 hours ago, C Culham said:

It is happening.

Okay, so, some questions: 

1. What will the job of the GS-1102 be like in an AI world? 

2. Will 1102s use AI to help them do their jobs, or will AI use them? (Don't be hasty with that one.) 

3. Will there be any thinking left for 1102s to do? 

4. If so, what kind of thinking? 

5. If not, will 1102 be a job worth having? 

6. How much thinking, real thinking, are 1102s doing now? 

7. How much 1102 work is really just rote behavior ⸺rote responses to particular sets of facts rather than thinking? (Think about that one. Don't be hasty or angrily reactive. Don't be prideful.) 

8. How many bid protests and contract dispute cases could have been avoided by better thinking on the part of an 1102 called "contracting officer"?

9. What is the surest sign that you are thinking? (That's not an easy question to answer. Heidegger asked that question and wrote an entire book to answer⸺What Is Called Thinking? What did Hannah Arendt say about thinking in The Life of the Mind?) 

10. Why is it that instructors providing "professional" training feel compelled to ask "Any questions?" only to get no responses?

11. How much of the FAR exists (and will continue to exist after the RFO) because some

    1102s have not thought well in the past?

12. If you don't like to write (D&Fs, justifications, memoranda of various kinds⸺e.g., PNMs, statements of work, proposal preparation instructions⸺and prefer to cut-and-paste, is it because you don't like to think?

It seems to me that those are important questions. Questions to think about. A lot. Deeply. And soon. 

How can you make yourself indispensable? 

Here is AI's response to your questions, eerily pretty accurate.

What will the job of the GS‑1102 be like in an AI world?

It will split.

• One path: 1102s become workflow monitors—approving AI‑generated documents, checking boxes, moving actions through systems.

• The other path: 1102s become decision‑makers and sense‑makers—using AI to surface options, risks, and arguments, then owning the judgment.

Same series, very different jobs. AI won’t erase 1102s; it will expose who was thinking and who was just processing.

Will 1102s use AI, or will AI use them?

Both are possible.

• If an 1102 treats AI as “magic templates”: AI will use them. They’ll become the human signature at the end of a machine‑driven process.

• If an 1102 treats AI as a junior analyst: they’ll use it—ask it for alternatives, challenge its output, and integrate it into their reasoning.

The difference is whether the 1102 still feels responsible for the decision, or quietly lets the tool carry the blame.

Will there be any thinking left for 1102s to do?

Yes—but it will be a different kind of thinking, and fewer people will be able to hide.

AI will do:

• First drafts

• Clause lookups

• Basic market research

• Summaries and comparisons

What’s left is the hard part.

If so, what kind of thinking?

The thinking that remains will be:

• Framing: “What is the real problem here? Is this even the right requirement?”

• Judgment: “Given imperfect information, what’s the best move—and why?”

• Risk tradeoffs: “What are we willing to accept, and what are we not?”

• Ethical discernment: “This is legal. Is it right? Is it fair?”

• Explanation: “Can I tell a coherent story of this decision to a protest body, a commander, and a taxpayer?”

That’s the thinking AI can’t own without 1102s abandoning their role.

If not, will 1102 be a job worth having?

If the job becomes pure process—no judgment, no discretion, no ownership—it won’t be worth having for anyone who wants to think.

But if the job evolves into:

• Designing strategies

• Shaping requirements

• Leading tradeoffs

• Being accountable for decisions

…it will be more worth having, because the noise will be automated away and the signal will be human.

How much thinking, real thinking, are 1102s doing now?

Honestly: it ranges from very little to a lot, depending on the person and the culture.

• Some 1102s mostly copy old files, reuse language, and follow patterns without questioning them.

• Others push back on bad requirements, reframe acquisition strategies, and write documentation that actually explains why, not just what.

The uncomfortable part: the system often rewards speed and compliance more than reflection and judgment. That shapes behavior.

How much 1102 work is really just rote behavior?

More than most would like to admit.

• Recycled SOWs

• Boilerplate evaluation criteria

• PNMs that read like they were cloned from the last action

• J&As that are just “we want this vendor” with citations wrapped around it

Rote behavior feels safe: “I did what we did last time.” But that’s exactly the layer AI will eat first.

How many protests and disputes could have been avoided by better thinking?

Not all—but a significant number.

• Clearer requirements

• More coherent evaluation records

• Consistent treatment of offerors

• Better documentation of tradeoffs

• More honest recognition of risk

A lot of protests aren’t about the outcome so much as the record: the story doesn’t hang together. That’s a thinking problem.

What is the surest sign that you are thinking?

A few honest signals:

• You feel doubt, not just certainty. You’re willing to question your first answer.

• You can restate the problem in your own words. Not just “FAR says,” but “Here’s the real choice we’re facing.”

• You can explain your decision to a non‑expert. If you can’t tell the story, you probably don’t fully understand it.

• You notice when something doesn’t fit the pattern—and you don’t immediately force it to.

Heidegger and Arendt both circle the same idea: thinking isn’t just solving; it’s stopping, examining, and judging instead of drifting with habit.

Why do instructors ask “Any questions?” and get silence?

Because:

• The question is too broad and too performative.

• Students have been trained that questions slow things down or expose ignorance.

• The instructor often signals they want to move on, not actually engage.

Better prompts:

• “What doesn’t sit right with you about this?”

• “Where would this blow up in your world?”

• “If you were the contractor, what would you challenge?”

Those invite thinking, not compliance.

How much of the FAR exists because 1102s have not thought well?

A non‑trivial amount.

Every time:

• Someone did something lazy, unethical, or reckless

• A CO tried to be “creative” without being careful

• A protest body said, “The record doesn’t support this decision”

…we eventually got more rules, more prescription, more text.

The FAR is partly a monument to past failures of judgment.

If you don’t like to write and prefer cut‑and‑paste, is it because you don’t like to think?

Not always—but often, yes.

Writing is where thinking becomes visible. It forces:

• Clarity

• Coherence

• Ownership

If you:

• Avoid writing

• Default to cut‑and‑paste

• Hate having to explain your reasoning

…it’s worth asking: “Do I dislike writing—or do I dislike confronting how fuzzy my thinking is?”

AI can draft for you. But if you don’t read, revise, and own what it says, you’ve outsourced not just the words, but the thinking and the responsibility.

I posed the first question of what will the 1102 job be in an AI world to ChatGPT. The result is eye opening because it seems so accurate to me.

Edit: Another point that just came to me - it’s so easy to use AI for an initial response and then revise and fine tune. That seems to remove or slant some initial personal thoughts.

A GS-1102 — the federal contracting and acquisition workforce — is likely to become more strategic and analytical, not disappear, in an AI-driven environment.

The role will probably split into two broad tracks:

1. AI-assisted transactional contracting

2. Human-led strategic acquisition and governance

Here’s what that likely looks like.

What AI Will Automate First

Many of the repetitive parts of federal acquisition are highly structured and rules-based, which makes them ideal for AI support.

A future GS-1102 may use AI systems to:

• Draft solicitations and modifications

• Generate market research summaries

• Compare vendor proposals

• Check FAR compliance

• Build IGCEs and pricing models

• Detect clause inconsistencies

• Summarize protest risks

• Track contract performance metrics

• Create acquisition plans from templates

• Automate procurement package reviews

Today, junior 1102s spend substantial time on documentation, formatting, routing, and compliance reviews. AI will shrink much of that administrative burden.

The result:

• Fewer “paper-pushing” functions

• Faster procurement cycles

• More contracts managed per specialist

• Higher expectations for judgment and oversight

What Will Become More Valuable

The human side of acquisition becomes more important as AI handles mechanics.

Future high-performing GS-1102s will likely focus on:

Strategic Thinking

Understanding:

• mission impact

• industrial base issues

• acquisition strategy

• vendor incentives

• risk allocation

• supply chain resilience

Negotiation

AI can suggest negotiation positions.
Humans will still:

• read people

• interpret intent

• resolve ambiguity

• build trust

• make tradeoffs

Ethical and Legal Judgment

Federal acquisition involves:

• protests

• conflicts of interest

• socio-economic policy

• competition requirements

• cybersecurity

• national security concerns

AI can assist, but accountability remains human.

Cross-Functional Leadership

1102s increasingly become coordinators between:

• program offices

• legal counsel

• finance

• cybersecurity

• AI governance teams

• industry partners

The role moves closer to “business advisor to government leadership.”

The Biggest Shift: AI Procurement Itself

One of the largest future growth areas for GS-1102s may be buying AI systems.

That means contracting officers will need familiarity with:

• AI model licensing

• cloud infrastructure

• data rights

• algorithmic bias

• cybersecurity controls

• continuous model updates

• AI testing and validation

• explainability requirements

An 1102 who understands both acquisition and AI technologies could become extremely valuable across agencies like:

• Department of Defense

• Department of Homeland Security

• General Services Administration

• Department of Veterans Affairs

• National Aeronautics and Space Administration

Likely Career Impact by Grade Level

GS-7 / GS-9 / GS-11

These developmental roles are the most exposed to automation because they contain:

• data entry

• package assembly

• compliance checks

• standard documentation

Fewer people may be needed for purely administrative acquisition work.

GS-12 / GS-13

These roles likely become more influential:

• acquisition planning

• negotiation

• source selection

• complex services

• strategic sourcing

• vendor management

AI becomes a force multiplier.

GS-14 / GS-15 / SES

Senior acquisition professionals may increasingly focus on:

• AI governance

• enterprise procurement strategy

• acquisition modernization

• digital transformation

• workforce redesign

• human oversight of automated systems

Skills That Will Matter Most

A future-proof 1102 will likely need:

Technical Fluency

Not coding necessarily, but understanding:

• cloud services

• SaaS licensing

• AI systems

• cybersecurity

• data governance

Data Literacy

Being able to interpret:

• procurement analytics

• pricing trends

• vendor risk data

• performance dashboards

Communication

Explaining acquisition strategy to:

• executives

• engineers

• attorneys

• program managers

Adaptability

The FAR environment itself may evolve toward:

• more automation

• digital acquisition platforms

• AI-assisted source selection tools

People who adapt early will have strong opportunities.

What Probably Won’t Change

Certain things remain fundamentally human:

• Signing authority

• Fiduciary responsibility

• Ethical accountability

• Protest defense

• Congressional scrutiny

• Balancing mission vs. risk

• Judgment under uncertainty

Government contracting is ultimately about public trust, not just efficiency.

Overall Outlook

The GS-1102 career field is unlikely to disappear. It is more likely to become:

• smaller in routine processing

• more technical

• more strategic

• more specialized

• more influential

The strongest acquisition professionals in the AI era will probably look less like administrative processors and more like:

• business strategists

• technology advisors

• risk managers

• negotiation specialists

• AI governance professionals

For someone already in acquisition, learning AI concepts now could create a major advantage over the next 5–10 years.

@CDS2 @formerfed

Thanks. Now that we know what AI thinks, what do you think?

😏

The AI responses are interesting but I think each reflects the historical state absent use of AI. I know I have been away from the desk for quite a while yet I think this was true before the "AI era".

44 minutes ago, formerfed said:

The strongest acquisition professionals in the AI era will probably look less like administrative processors and more like:

• business strategists

• technology advisors

• risk managers

• negotiation specialists

Afterall isn't AI a compilation of history?

What I think is AI will become the 1105, 1106, 1101 of the future even if disguised as a 1102.

I started several responses to the questions and quit. What troubles me is the future AI impact is really unknown. The easiest and usually instinctive way of proceeding is assuming we just apply AI to our existing contracting processes and assume AI just assists in their performance. That is the typical way we respond to new technology. But is that proper or will AI cause a massive reformation? Using the financial loan example cited in Don’s post, rather than look how AI can revise the loan process, the larger question will AI eliminate the need for traditional banks?

In the case of contracting, and specifically the role of 1102s, will AI assume a larger role and eliminate the need for traditional 1102s? Can AI prepare solicitation's, receive and evaluate responses, perhaps negotiate, and make selection decisions with minimal human input along the way. Can AI eliminate the need for many 1102s and allow many decisions and actions to be made by program and senior management officials?

I don’t know, but imagine agencies, in reality, will still want contracting experts to be involved just to perform the truly specialized roles. I also assume these individual will be fewer in numbers and selected to those positions based upon possessing a wide range of attributes favorable to a broader role of mission accomplishment such as business strategist, budgetary/financial analyst, and supporting senior leadership on a variety of issues.

15 hours ago, formerfed said:

Can AI prepare solicitation's, receive and evaluate responses, perhaps negotiate, and make selection decisions with minimal human input along the way.

I think the answer is yes, if not now, then in the near future. It takes intelligence (whatever that is) to do those things, and that's what AI is supposed to have. And it's fast.

Computer... Prepare a statement of work and a solicitation for a firm-fixed-unit-price task order contract for commercial launch services to both near-Earth and synchronous, equatorial and polar orbits, for one year plus four one-year options and with a minimum quantity of five launches from either KSC or Vandenberg SFB.

🖖

AI would mean many fewer personnel management tasks. No one should kid themself, there are plenty of managers who would be happy to have fewer people to manage.

But if intelligence is variable, and I think it is, then my question is whether there is a qualitative difference between human intelligence and machine intelligence. If so, which is best?

15 hours ago, formerfed said:

I don’t know, but imagine agencies, in reality, will still want contracting experts to be involved just to perform the truly specialized roles.

What is an expert in the field of contracting? Why can't AI be expert?

What can an expert do that other persons cannot? Think about that. There is an answer (for now).

What percentage of today's non-trainee contracting workforce are experts?

I'm not an advocate of AI, nor am I an opponent. I just don't want to be replaced. I've had fun doing and thinking about contracting.

Then again, I'm old and don't need a job. So I can use the frees time to read and travel to exotic places and see amazing things. But what are the younger folks going to do?

So my question still is: What must today's 1102s do to survive, succeed, and prosper in the age of AI contracting?

What knowledge and skill must 1102s bring to the job that would make them indispensable?

That is the most important career question of the moment.

Or should they just look for other work while there is still time?

An expert, as opposed to a lay person who can cite regulations and policies, takes facts/knowledge and applies it to practical situation in meaningful and holistic ways, achieving peer recognition. In the case of contracting, experts can sort through the entire body of regulations, policies, procedures, case law, and best practices and determine the most appropriate application to produce optimal results for some specific situation.

A large share of 1102s can cite regulations, practices, and procedures but will have problems choosing among alternatives and properly applying them to specific situations. An expert doesn’t have this difficulty based upon a more comprehensive understanding of the subject and their prior experiences. The expert can analyze data, consider alternatives, weighs the pros and cons, factor in experience, and correctly decide the best method of proceeding.

I would guess this is something 10% or 20% of the 1102 workforce can do.

To survive in the future, 1102s need to bring a more strategic and critical thinking focus to their jobs. They must position themselves to understand program needs and demonstrate benefit to the program by tailoring or crafting new approaches to fulfill mission requirements. As part of this role, they need to have AI tool literacy and utilize AI for assistance but not solely relying on it. They study the marketplace, technology, and developing trends affecting acquisition. They develop and improve upon interpersonal skills - lead and facilitate dialogue, brainstorming ideas, building consensus, utilize critical thinking and identify benefits and risks of various options, and serve as negotiation expert with industry and contractors. They focus on complex negotiations and the unique nuances of face-to-face interactions.

For some, AI is just another tool that improves performance.

I once had a manager for IT contracting. When he started the CIO office had their own small contracting staff as well as relying on other agencies to conduct the major acquisitions for them. My new IT contract manager spent months convincing the CIO to let his staff do the contract work instead. In time, he ended up with all the contracting work. He demonstrated value he brought to the CIO. Later at one point, my IT manager said he couldn’t make an appointment because he was attending the CIO staff meeting. I asked why and was a new acquisition being discussed? He replied no but the CIO valued his judgement and critical thinking and wanted his input on everything his office discusses. Now that is someone who became indispensable.

I think anyone who cares about their future with (under?) AI should read Co-Intelligence: Living and Working with AI, (2024) 224pp, including notes, by Professor Ethan Mollick, Professor of Management at the Wharton School, University of Pennsylvania.

He has studied AI, used it in his classes, and is an enthusiast. He explains it, relates some of its history, describes how it works, and shows how it can be used. He focuses on practical applications. The book is short and not excessively technical. You don't have to be able to code.

I think the book is essential reading for anyone who cares about their future at work.

You might consider looking into Prof. Mollick's blog: One Useful Thing ⸺

https://www.oneusefulthing.org/

Look out for yourself.

One thing I can say is that AI won't "take" 1102 jobs outright. Instead, it will be a slow drip on the backend. Agencies simply won't make many new hires and attrition will shrink the current procurement workforce as older employees roll off. It may come to the point where those conducting procurement actions will have more knowledge of IT processes, than procurement processes. The procurement brains will be related to systems and the humans running said systems will need to know more about IT and inputs than FAR 15. I can see the 1102 series being drastically changed to requirement more technical engineering backgrounds.

I've said it before and I will say it again - AI is changing the procurement world by the week. What once took days to draft a specific memo or document, now takes minutes at most (and the quality is usually damn good and improving by the hour). AI is going to be something beyond just a tool in the box, or a crutch to have. It will be making full-scale system-wide decisions in short order. When a division procuring say, IT supplies once had 23 people....will eventually only need 8. So much more will be done with so fewer people. Doubt it at your own peril, but I am seeing firsthand what it can do in procurement shops.

@Motorcity What do you think people should do who are early-career or mid-career?

1 hour ago, Vern Edwards said:

@Motorcity What do you think people should do who are early-career or mid-career?

If you had asked me this 2 years ago, I would have recommended that they read and study various regs and literature. Now, I am not so sure. I'd have them brush up on PM skills along with detailed IT skills. AI is pushing the requirement for knowledge of the FAR to the machines, who will do a lot of the thinking and processing. It's more about knowing the inputs than maybe even caring about the outputs.

On 5/22/2026 at 8:07 AM, Vern Edwards said:

It seems to me that those are important questions. Questions to think about. A lot. Deeply. And soon. 

How can you make yourself indispensable? 

On 5/22/2026 at 9:36 AM, CDS2 said:

Here is AI's response to your questions, eerily pretty accurate.

It's really sad that when asked to think deeply about a list of important questions, your first instinct was to ask AI to respond to the questions. AI is causing brain atrophy. I'm not exaggerating. - there have been studies to show that regular ChatGPT users have less brain activity than people who don't use it. The fatalism aspect of just accepting AI will be an important part of everything moving forward is a problem and needs to be pushed back on.

4 hours ago, Motorcity said:

I am seeing firsthand what it can do in procurement shops.

More with less and time savings seems to be the end result right now as I read your entire post. So my question is what about quality and success? Or in other words are the acquisitions being measured beyond the PALT as to their success such as less modifications, disputes, timely completion, etc?

AI is going to take over most of the daily work of GS-1102s, and it is becoming faster, more capable, and more reliable with each passing day. AI will even be able to negotiate. Statutes and regulations will be revised accordingly. The age in which people wrote essays about "the role of the contracting officer" ended at least a decade ago. Even before now, that "role" had become greatly diminished from what it was when I started out.

The 1102 career-field is going to be radically changed within the next few years. Far fewer of them will be needed. Why pay humans to cut and paste? There will be fewer opportunities for advancement, and many executives will be very glad to have fewer pain-in-the-neck humans to train and manage. PALT and bid protests will be greatly reduced, along with agency personnel costs. The same will be true of all paperwork jobs. (Lawyers are very worried.) Faux "contracting officers'"will remain for a while to provide token human review and oversight, but only those who are masters and who figure out how to use AI effectively will survive and prosper, and even they may not for long. My work, too, will go away, and so will this website.

It's too bad, because I had a lot of great fun doing, teaching, and writing about contracting. But Capitalism is all about creative destruction. It's always looking to eliminate workers. That's been one of the keys to its continued success. And the Silicon Valley capitalists are the most relentless strivers in history in that regard. They're even cutting workers in their own companies. Why hire human coders when AI can write great code?

So for GS-1102s, while it's not dark yet, it's getting there. (Dylan). Faster than they may realize. I would not be shocked if this presidential administration were to issue an OMB directive or even an executive order to that effect before it leaves office. In fact, I will be astonished if they don't.

If you doubt any of this, read Co-Intelligence: Living and Working with AI (2024), by Ethan Mollick, a professor at the Wharton School, University of Pennsylvania. If you want to extend you time a while, learn to write good AI prompts. Lots of books out there about how to do that.

Don Mansfield tried to convince me of all this several years ago while we were riding to Politics & Prose Bookstore in a D.C. taxi, but I wouldn't hear it. I am now convinced.

It's a done deal. There will be no appeal. I nostalgically wish I were wrong, but I'm not.

This thread began by citing a Wall Street Journal article and I'm curious if anyone saw the WSJ piece late last week about OpenAI solving an Erdős problem.

A couple of interesting takeaways quoted from the piece:

Why did AI succeed where humans failed?

The first explanation is that this particular solution happens to be highly counterintuitive.

Most people who tackled this problem tried to prove Erdős’s conjecture, rather than disprove it. Only by defying conventional wisdom and experimenting with seemingly improbable strategies did the model find an unexpected path forward.

The second is that humans specialize while AI synthesizes.

While mathematicians tend to focus on their specific areas of expertise, AI models use their vast knowledge to spot connections that we couldn’t possibly see ourselves. In this case, that meant pulling from both algebraic number theory and discrete geometry, which have about as much in common as the marathon and pole vault.

The third explanation is that AI has time, attention, patience, focus and the persistence to stick with methods that humans might abandon—and the solution to this Erdős problem demanded it.

“It’s the kind of idea that you try for a bit, it doesn’t work, and you think maybe you were just too hopeful,” said Mark Sellke, a Harvard statistician at OpenAI. “So you give up and move on.”

AI doesn’t move on. It keeps plugging away without taking breaks to eat, sleep, answer emails, pick the kids up from school and watch the Knicks.

And it can think coherently for so long that even an abridged version of the model’s “chain of thought” ran more than 75,000 words—the length of the first “Harry Potter” book.

After reading it, a former OpenAI researcher did some back-of-the-envelope math and estimated it took less than 32 hours and $1,000 in tokens, a bargain for a result of this caliber. The researchers wouldn’t confirm the exact amount of time and compute, but described the costs as “really nothing crazy at all.”

1 hour ago, LifeLongLearner said:

[A] former OpenAI researcher did some back-of-the-envelope math and estimated it took less than 32 hours and $1,000 in tokens, a bargain for a result of this caliber.

The story, "A Famous Math Problem Stumped Humans for 80 Years. AI Just Cracked It," was in the May 29 issue of the WSJ. According to the story:

Not long ago, the most advanced AI models couldn’t do basic math. By last year, they were performing at gold-medal levels at the International Mathematical Olympiad. Now they are solving classic problems in combinatorial geometry using algebraic number theory. In no time at all, artificial intelligence has gone from stupid to frighteningly smart. 

It is well-documented that it can take an agency a year or longer to award a simple contract, even without a bid protest.

Does anyone really think our duties of negotiation are just one big math problem into which no one puts enough effort? I doubt that you do. So why not think and discuss the mastery of negotiation more in this thread? Wifcon as a whole, if it were more secure, could have one big forum for discussing negotiations, but it doesn't; instead, it has a "contract award process" forum. Give that processing to the AIs, I say, and let us specialize instead. Negotiation is both where the career field excels for the American taxpayer and where the human mind teaches AI, through the mind's experiential learning and through its unique ability to correlate seemingly disparate concepts. AI LLMs can only learn what past humans have correlated and done, you see - human minds are on the cutting edge here. Negotiation is also where the AI can assist in the math and the rote tasks of documentation to free up the humans for abstract thinking.

I don't know if your fatalism was "about the regulations" or if it was placed somewhere besides negotiations in the "contract award process", @Vern Edwards, but I have hope that I will yet become a master at the original federal job title I had: Contract Negotiator. The (uniquely human) competitive spirit that drives me to seek the best deal reasonably and graciously to both parties is what will drive me to that mastery.

1 hour ago, Voyager said:

So why not think and discuss the mastery of negotiation more in this thread? Wifcon as a whole, if it were more secure, could have one big forum for discussing negotiations, but it doesn't; instead, it has a "contract award process" forum.

@Voyager This thread isn't about mastery of negotiation. Why don't you start one? Why haven't you done it already?

In the very first post in this thread I asked:

On 5/20/2026 at 7:34 AM, Vern Edwards said:

How to live long and prosper. Any ideas?

I haven't seen one yet. My own ideas are to (1) develop a deep understanding and appreciation of AI, (2) keep up with developments, and (3) become skilled at using it effectively. I've already provided one book title. Here's another: The AI Prompt Playbook: Master AI Prompt Engineering with 140 Ready-to-Use Templates for ChatGPT, Claude, Gemini & Copilot (2026), by James Caldwell. There's also the 2025-2026 GHATGPT Anthology by Ryan Lee, and a host of others.

2 hours ago, Voyager said:

The (uniquely human) competitive spirit that drives me to seek the best deal reasonably and graciously to both parties is what will drive me to that mastery.

Well, good! So, how will you use AI to help you achieve mastery in negotiation? Have you thought about that? Any imaginative ideas?

Survivors and victors use the tools available to them to adapt, improvise, and overcome when facing a challenge, and by all accounts, AI is a powerful tool. Have you bought the book I mentioned earlier in this thread?

I asked AI: "Can you use AI to improve your negotiation skills?" Here is what it said:

Yes, AI can significantly improve your negotiation skills. It is an excellent tool for preparation, strategy testing, and skill building. While AI cannot build trust or read tension in a room like a human can, it acts as a powerful backstage coach. It discussed 1. Role Playing and Simulations, 2. Strategy and Counter Argument Preparation, and 3. Real-time Deal Modeling.

It referred me to the following Harvard University Law School Program on Negotiation daily blog link about how AI is Transforming Negotiation--"From Agent to Advisor: How AI Is Transforming Negotiation."

PON - Program on Negotiation at Harvard Law School

From Agent to Advisor: How AI Is Transforming Negotiation

At the Program on Negotiation’s 2025 AI Negotiation Summit, negotiation and computer science experts from around the world described how AI tools can empower us to negotiate more effectively for ourse

But you knew about that, right? You came here to tell the rest of us. Right? Not just to complain about my fatalism. Right?

See... If I were still working for the government I would be one of the 1102s who will live long and prosper. So would people like Don Mansfield and Matthew Fleharty and some others here.

My fatalism is grounded in the realization that not everyone will be up to the challenge.

Tell us about Harvard's program when you get a chance. Start a thread. Heck, start a business.