AI-Based Prose Programming for Subject Matter Experts: Will This Work?

AI-Based Prose Programming for Subject Matter Experts: Will This Work?

Key Takeaways

  • Recent advances in prose-to-code generation via Large Language Models (LLMs) will make it practical for non-programmers to “program in prose” for practically useful program complexities, a long-standing dream of computer scientists and subject-matter experts alike.
  • Assuming that correctness of the code and explainability of the results remain important, testing the code will still have to be done using more traditional approaches. Hence, the non-programmers must understand the notion of testing and coverage.
  • Program understanding, visualization, exploration, and simulation will become even more relevant in the future to illustrate what the generated program does to subject matter experts.
  • There is a strong synergy with very high-level programming languages and domain-specific languages (DSLs) because the to-be-generated programs are shorter (and less error prone) and more directly aligned with the execution semantics (and therefore easier to understand).
  • I think it is still an open question how far the approach scales and how integrated tools will look that exploit both LLMs’ “prose magic” and more traditional ways of computing. I illustrate this with an open-source demonstrator implemented in JetBrains MPS.



As a consequence of AI, machine learning, neural networks, and in particular Large Language Models (LLMs) like ChatGPT, there’s a discussion about the future of programming. There are mainly two areas. One focuses on how AI can help developers code more efficiently. We have probably all asked ChatGPT to generate small-ish fragments of code from prose descriptions and pasted them into whatever larger program we were developing. Or used Github Copilot directly in our IDEs.

This works quite well because, as programmers, we can verify that the code makes sense just by looking at it or trying it out in a “safe” environment. Eventually (or even in advance), we write tests to validate that the generated code works in all relevant scenarios. And the AI-generated code doesn’t even have to be completely correct because it is useful to developers if it reaches 80% correctness. Just like when we look up things on Stackoverflow, it can serve as an inspiration/outline/guidance/hint to allow the programmer to finish the job manually. I think it is indisputable that this use of AI provides value to developers.

The second discussion area is whether this will enable non-programmers to instruct computers. The idea is that they just write a prompt, and the AI generates code that makes the machine do whatever they intended. The key difference to the previous scenario is that the inherent safeguards against generated nonsense aren’t there, at least not obviously.

A non-programmer user can’t necessarily look at the code and check it for plausibility, they can’t necessarily bring a generated 80% solution to 100%, and they don’t necessarily write tests. So will this approach work, and how must languages and tools change to make it work? This is the focus of this article.

Why not use AI directly?

You might ask: why generate programs in the first place? Why don’t we just use a general-purpose AI

Read More

The Chip Titan Whose Life’s Work Is at the Center of a Tech Cold War

The Chip Titan Whose Life’s Work Is at the Center of a Tech Cold War

In a wood-paneled office overlooking Taipei and the jungle-covered mountains that surround the Taiwanese capital, Morris Chang recently pulled out an old book stamped with technicolor patterns.

It was titled “Introduction to VLSI Systems,” a graduate-level textbook describing the intricacies of computer chip design. Mr. Chang, 92, held it up with reverence.

“I want to show you the date of this book, 1980,” he said. The timing was important, he added, as it was “the earliest piece” in a puzzle that came together for him — altering not only his career but also the course of the global electronics industry.

The insight that Mr. Chang gained from the textbook was deceptively simple: the idea that microchips, which act as the brains of computers, could be designed in one place but manufactured somewhere else. The notion went against the semiconductor industry’s standard practice at the time.

So at the age of 54, when many people begin thinking more about retirement, Mr. Chang instead put himself on a path to turn his insight into a reality. The engineer left his adopted country, the United States, and moved to Taiwan where he founded Taiwan Semiconductor Manufacturing Company, or TSMC. The company does not design chips, but it has become the world’s biggest manufacturer of cutting-edge microprocessors for customers including Apple and Nvidia.

Today, the company that partially exists because of a textbook is a $500 billion juggernaut that has put the most advanced chips in iPhones, cars, supercomputers and fighter jets. So critical are its airplane-hangar-size chip factories, called fabs, that the United States, Japan and Europe have courted TSMC to build them in their neck of the woods. Over the past decade, China has also invested hundreds of billions of dollars to recreate what TSMC has done.

Mr. Chang’s unlikely entrepreneurial journey helped Taiwan become an economic giant, restructured the way the electronics industry worked and ultimately charted a new geopolitical reality in which a linchpin of global economic growth lies in one of the world’s most volatile spots.

That has thrust Mr. Chang, and the company he created, into the spotlight. And at the twilight of his career, a man who has preferred to remain in the shadows reflected on what he has built and what it means to no longer be able to stay under the radar.

“It doesn’t make me feel particularly good,” said Mr. Chang, who retired in 2018 but still appears at TSMC events. “I would rather stay relatively unknown.”

Over a recent three-hour discussion in his office, Mr. Chang made it clear that he identifies as American — he obtained his U.S. citizenship in 1962 — at a time when the company he founded is at the center of a technological Cold War between the United States and China. Even as the rivalry for tech leadership intensifies, he does not give China much of a chance for semiconductor supremacy.

“We control all the choke points,” Mr. Chang said, referring collectively to the United States and its

Read More

A fifth of US employees have work opportunities with ‘high exposure’ to AI

A fifth of US employees have work opportunities with ‘high exposure’ to AI

About a person-in-five U.S. employees have positions with essential tasks that are extra likely to be aided or changed by AI, according to a modern report from Pew Exploration Heart.

The results, dependent on an investigation of federal details, identified that work opportunities that rely on analytical capabilities like important contemplating, writing, science and math tend to be “additional exposed” to the rising technological know-how. Apparently, staff in industries additional exposed to AI are far more very likely to say they assume it will help relatively than damage their work opportunities, in accordance to a Pew survey.

“Employees who are more familiar with AI appear to be looking at extra positive aspects than harm,” explained Rakesh Kochhar, a senior researcher at the nonpartisan imagine tank who authored the report.

The report pointed out that it’s unclear how lots of employment are at risk due to AI, although some results recommend work opportunities are currently getting lost to the technology. AI contributed to approximately 4,000 task cuts in Might, in accordance to a report from Challenger, Gray & Xmas. 

A fifth of US employees have work opportunities with ‘high exposure’ to AI

Which positions are most at-possibility because of to AI?

U.S. careers likely to have significant, medium and low publicity to AI involve:

Superior publicity:

  • Spending budget analysts
  • Information entry keyers
  • Tax preparers
  • Complex writers
  • Web builders

Medium publicity:

  • Chief executives
  • Veterinarians
  • Interior designers
  • Fundraisers
  • Sales supervisors

Small exposure:

  • Barbers
  • Youngster treatment workers
  • Dishwashers
  • Firefighters
  • Pipelayers

In sum, about 19% of U.S. employees had been in employment most uncovered to AI final calendar year, whilst an even better share (23%) experienced work considered least uncovered.  

It is really not crystal clear how a lot of work will be displaced by AI. A March report from Goldman Sachs found AI could substitute up to 25% of current function, with about two-thirds of work opportunities uncovered to “some diploma” of automation.

But researchers note that displacements pursuing the emergence of new technological know-how have typically been offset by the development of new jobs, with census details suggesting that about 60% of employees today are utilized in positions that did not exist in 1940.

Which staff members are most at hazard?

Pew uncovered that women, Asian, higher education-educated and bigger-compensated staff are much more uncovered to AI. 

Kochhar explained this is due to the fact of the kinds of positions held by distinctive demographics: adult males tend to hold a lot more jobs requiring bodily labor like design, for instance.

“So at the second, they have fewer exposure to AI,” Kochhar mentioned. “Which is not to say AI could not direct to smarter robots that can do it all, also. Which is not a thing we appeared into.”

According to the report:

  • Personnel with a bachelor’s diploma (27%) are far more most likely than people with only a significant faculty diploma (12%) to maintain a job with the most publicity to AI.
  • Girls (21%) are far more probably than gentlemen (17%) to have jobs with the most publicity to AI.
  • Black (15%) and Hispanic
Read More

Farming At Your Fingertips: How Technology is Modifying Agriculture Work

Farming At Your Fingertips: How Technology is Modifying Agriculture Work

About five yrs back, David Wallace expended the summer months on his family’s potato farm in the Skagit Valley place of Washington State. Wallace experienced grown up on the farm, but he remaining to go after chemistry, sooner or later doing the job as a facts scientist. But he often felt a pull back again to his roots. 

During his pay a visit to, Wallace listened as his father complained about the farm’s irrigation techniques. They ended up prone to mistakes, losing enormous amounts of drinking water and time. Every single time there was an situation with the movement amount or the center line, the crew would have to generate around the subject to obtain the correct stage of difficulties and then determine out how to repair it. The most important issue, the elder Wallace explained to his son, was that there was no way to see what was going on with the irrigation system remotely. Following a swift research, Wallace identified that there was not a enough checking and manage program obtainable on the market. So, he designed a single. And just like that, FarmHQ was born. 

The smaller unit attaches to a central level on the irrigation system and makes use of cloud-based software program to keep track of irrigation reels and pumps. Wallace spent months tinkering with his code before testing it out across his farm and with a several buddies. Rapidly, phrase distribute, and other farms desired in on the time-conserving product. Now, FarmHQ is in its 3rd calendar year of development and utilized on about 30 farms across the Pacific Northwest, with ideas to develop tenfold this 12 months. Wallace in no way returned to his previous knowledge scientist task. 

Technological know-how this sort of as FarmHQ aims to enable farmers turn out to be a lot more economical, conserving time and funds. Wallace statements implementing his process can web up to a 1,500-percent return on financial investment about a time. “These devices pump wherever from 250 to 400 gallons of drinking water for each minute over a incredibly smaller location of land,” he claims. “We can work out specifically how much pumping time and how a lot h2o is saved as a end result of having our technique on board. We also know incredibly accurately how significantly driving time we’re conserving farmers, due to the fact we know the place of every 1 of their parts of machines, and we know wherever their home base is. So each and every time they open up up that app to check out on that piece of tools, we’ve basically saved them a vacation to the area.”

[RELATED: At This Farm, Data Is the Most Important Crop]

Some of the technology embraced by farmers has a bodily component, like FarmHQ. Other people are only software and apps that aid acquire the multitude of data that arrives from farming, and it makes the procedure more collaborative. One particular these types of merchandise is Agworld, basically a

Read More

I Eventually Located a Work Laptop or computer Set up That’s Basically Fantastic

I Eventually Located a Work Laptop or computer Set up That’s Basically Fantastic

From my early times in match growth to my existing composing gig, I’ve been working at a desk for far more than 25 many years. Screens have always commanded my notice. Items experienced improved enormously due to the fact two chunky beige CRTs dominated my desk, but I continue to hadn’t found the perfect monitor setup—until now.

I’m divulging this info to you for absolutely free (you’re welcome). If you search for the great household office set up, you require a curved ultrawide monitor with a 2nd monitor on the facet in portrait orientation. It’s the best combination for writing and gaming, with adequate actual estate for each process.

If you invest in a little something using inbound links in our stories, we might gain a fee. This can help support our journalism. Study extra.

Twin-Display screen Need

Photograph: Simon Hill

As a kid, I longed for a battle station with a number of monitors, but that kind of set up was uncommon back then, identified only in the offices of obsessive builders (or the lairs of supervillains). My very first job after college was as a match tester. The indignity of regularly doing the job 16-hour shifts in a transformed closet with wires dangling from the ceiling was assuaged by the truth that I had two screens—one to engage in by means of the recreation and the other to log bugs.

The geeky thrill of that dual-screen arrangement sparked a lifelong lookup for the excellent combo. As I sophisticated from play-screening to match design and style, my set up gradually enhanced. But funds was limited at get the job done and home. For years, I experienced two mismatched screens, balanced precariously on a pile of textbooks. With two screens, you can perform on one particular and examination on the other, exploration on one particular and publish on the other, and so on. 

But it is a flawed setup for gaming. You can extend the match throughout both of those screens, but the keep track of bezels in the center ruin the knowledge. So alternatively I would participate in game titles on a person and use the second for new music, podcasts, and messaging. That setup was absolutely wonderful, but my drive for symmetry designed the mismatched sizing irritating. 

Even after I upgraded to equivalent monitors, continually transferring in between them intended I was generally dealing with a little bit off to the facet. Not a issue for a supple youngster, but above the many years I created cricks and aches that led to serious back again pain.

Three’s a Group

When I acquired my 1st flat-display screen Tv set, I couldn’t hold out to try it out with my Personal computer. The rewards of one particular substantial display had been instant. I could look straight on—no far more cricks in my neck. I rarely dropped keep track of of my mouse pointer, and the major screen was a dream to play games on. It was also neater and

Read More