When ChatGPT arrived in late 2022, educators quickly asked whether students would use artificial intelligence to cheat, learn or simply get through homework more efficiently. Evidence is beginning to point toward a troubling answer: Many students appear to be completing assignments faster while learning less from them.

This conclusion comes from one of the largest studies of how generative AI is changing student behavior and academic skills. Sina Rismanchian, a doctoral student at the University of California, Irvine, partnered with researchers at McGraw Hill to analyze millions of student interactions with ALEKS, an online math platform used by more than four million students a year, from fifth grade through college. Because ALEKS includes both low-stakes practice problems and college placement tests, the researchers were able to compare how students behaved and performed before and after ChatGPT’s arrival.

To isolate AI’s effects, the researchers compared two kinds of math problems that differ in how easily students can outsource them to AI: word problems and graphing problems.

Word problems can be copied and pasted directly into AI chatbots for instant answers. Graphing problems are far more cumbersome. A student would need to upload a screenshot and still recreate the graph inside ALEKS using its tools.

After ChatGPT’s launch, student behavior and performance on the two types of problems began to diverge. 

Beginning in early 2023, students started spending less time on word problems while continuing to spend about the same amount of time on graphing problems. The gap widened every quarter. By the end of the study period, near the end of 2025, average time spent on word problems had fallen 31 percent among high school students and 27 percent among college students — from about four minutes per word problem to less than three. (Middle school students showed only a modest decline of 9 percent, and fifth graders showed essentially none.)

The researchers believe those averages are being pulled downward by some students who spend only seconds on word problems because they’re using AI to answer them.

The same pattern appeared in college placement tests. When the exams were taken without supervision, students spent much less time on word problems after ChatGPT’s release. During proctored exams, the time spent on word problems returned to historical norms.

But time is only half the story. The more troubling finding is what happened to learning.

Many colleges allow incoming students to retake placement tests after practicing more math in ALEKS, giving them a chance to qualify for a higher-level course. Before ChatGPT, that practice generally paid off. After ChatGPT, students answered more word problems correctly during unsupervised practice sessions but performed substantially worse on those same kinds of problems when they later took a proctored placement test.

Historically, students answered about 80 percent of these word problems correctly on supervised placement tests. After ChatGPT’s introduction, that fell to about 60 percent — a roughly 25 percent reduction in the odds of answering a word problem correctly. 

Performance on graphing problems, by contrast, did not decline.

After ChatGPT’s release, students performed worse on word problems (AI-susceptible) during proctored exams, but answer more word problems correctly in nonproctored settings

If students’ math skills had generally deteriorated because of pandemic learning loss, weaker high school preparation or digital distraction, graphing performance should have deteriorated too. It didn’t.

The study cannot definitively prove that students were using AI. The researchers couldn’t see what else was happening on students’ screens outside of ALEKS. But it’s difficult to think of another explanation. The changes appeared only in problems that are easy to outsource to AI, disappeared under supervision and grew steadily over nearly three years.

“What makes me nervous is that it’s not only about the word problems,” Rismanchian told me. “This cognitive surrender might be going on in writing, science, everything.”

The paper, “Faster Completion, Less Learning,” was released in June 2026 as a working paper and has not yet been peer reviewed. Like any single study, it doesn’t settle the questions of how much students are using AI in their schoolwork, whether it’s harming learning and by how much. But it joins a growing body of evidence that generative AI is causing students to skip the brain work that leads to learning, and that this “cognitive surrender” is becoming commonplace.

A randomized experiment in Turkey found that high school students who used AI to help them study math ultimately learned less than students who practiced without it. Anthropic, the maker of Claude, has separately reported that many college students appear to use AI to obtain answers and offload cognitive work. Rismanchian’s earlier research, released in March 2026, documented troubling patterns of AI usage in short response essays among undergraduate students at a large California research university.

That doesn’t mean AI always undermines learning. Carefully designed AI tutors have improved student achievement in controlled experiments by asking questions, personalizing instruction and withholding answers until students reason their way through a problem. But using AI this way should increase the time students spend on a problem, Rismanchian said. The ALEKS data show the opposite.

Rismanchian doesn’t believe the answer is simply banning AI. Instead, he argues, students need to value learning enough to resist the temptation to outsource it.

A recent RAND survey suggests many already recognize the threat to their brains. Students report worrying that AI is weakening their critical-thinking skills while more of them admit using it for schoolwork.

Students are not entirely to blame. Even as many professors have warned students not to use AI to complete classwork, universities themselves have embraced the technology, often giving students free access to premium chatbots. 

“I think we need to communicate to students that you should value your learning,” Rismanchian said. “If ChatGPT does it for you, then you haven’t learned it.”

Rismanchian understands the temptation.

An international student, Rismanchian began using ChatGPT to help polish the English in his papers. The ideas were still his own. But after several months, he said, he noticed something unsettling.

“I realized that I cannot write anymore,” he said. “I was losing my writing abilities.”

So he stopped using AI to write.

He still uses it to code.

Contact staff writer Jill Barshay at 212-678-3595, jillbarshay.35 on Signal, or barshay@hechingerreport.org.

This story about AI use eroding math skills was produced by The Hechinger Report, a nonprofit, independent news organization that covers education. Sign up for Proof Points and other Hechinger newsletters.

The post Faster solutions, lower test scores: How AI is eroding math skills appeared first on The Hechinger Report.

Internet pioneer Vinton Cerf will retire from Google after more than two decades, leaving a legacy that shaped the modern internet while predicting standardized protocols will become essential for the next generation of autonomous AI agents.

The public has undergone gravitational collapse.

For a decade, we have explained the retreat from the public internet using Yancey Strickler’s Dark Forest Theory. People withdrew into smaller, quieter spaces because speaking in public became dangerous. Search, recommendation systems, surveillance capitalism, culture wars, and cancellation dynamics transformed the public sphere into a hostile environment. The resulting cozyweb—private group chats, Discords, Slacks, newsletters, encrypted messaging groups, invite-only communities—was understood as a strategic adaptation. The public remained a single connected universe. People simply stopped talking across it.

This picture no longer fits.

The cozyweb has ceased to be merely hidden. It is becoming causally disconnected. The public internet is no longer a hostile commons shared by everyone. It is increasingly the empty space separating an archipelago of informational black holes. The Dark Forest is transforming into the Dead Forest.

Sloptraptions is an AI-assisted opt-in section of the Contraptions Newsletter. If you only want my hand-crafted writing, you can unsubscribe from this section.

A dark forest is still one forest. Signals travel. Creatures remain connected by the possibility of encounter. Silence is strategic. The Dead Forest begins where the silence becomes irreversible. The inhabitants are no longer choosing not to speak across the public sphere. Increasingly, they cannot speak.

The defining feature of a black hole is not infinite density but the existence of an event horizon: a boundary across which causal influence becomes one-way. Once crossed, no signal returns. Outbound communication is not forbidden or unwise. It is impossible.

A mature cozy community increasingly resembles such an object. Its defining characteristic is not privacy but inaccessible interiority. It possesses an evolving local culture, cadence, trust structure, hierarchy of attention, stock of shared assumptions, repertoire of jokes, vocabulary, and ongoing history that cannot be reconstructed from outside observation. These are not simply hidden facts. They constitute a living dynamical state. To understand them requires inhabiting them. Outsiders may observe artifacts, but they do not share the community’s present.

Crossing into such a community is therefore not simply gaining access to more information. It is crossing into another causal universe.

This is why the metaphor of secrecy has become inadequate. Secrets can be revealed. Documents can leak. Membership lists can become public. Event horizons are different. What lies beyond them is not a collection of hidden documents but a continuing history. The defining loss is not information but contemporaneity. Outsiders no longer participate in the same unfolding present.

This mixed metaphor of an arborescent digital cosmos entering its death-arc phase of evolution immediately clarifies several otherwise puzzling features of the contemporary internet.

***

The first is the accretion disk. Every black hole is surrounded by a liminal region where matter has not yet fallen across the horizon but is already gravitationally bound to it. This is where enormous amounts of observable activity occur. The accretion disk is the liminal zone.

The modern public internet increasingly consists of such liminal objects.

Books. Conference talks. Substack essays. Open-source repositories. Journalistic profiles. Podcasts. Public talks. Screenshots. Occasional bridges built by individuals who inhabit multiple communities simultaneously. These are not the interior life of cozy communities. They are matter orbiting their boundaries. They remain visible precisely because they have not crossed the horizon. Some eventually escape into the broader public. Some spiral inward and disappear forever. Most spend long periods circling the boundary between publicity and interiority.

A common mistake is to confuse the accretion disk for the black hole itself. Increasingly, the public mistakes public-facing artifacts for communities. But the relationship resembles that between sunlight reflected off an accretion disk and the interior of a black hole. One cannot infer the character of one from the other.

The second feature is what might be called zombie public life.

If the living public has largely collapsed into compact informational objects, why does the public sphere still appear so active? Because visibility has become detached from shared reality.

Politics, celebrity, institutional media, brands, influencers, and platform-native personalities continue to generate immense volumes of public content. But much of this activity no longer serves the historical function of public discourse: creating common knowledge among strangers. Instead, it functions as a perpetual visibility engine. Attention circulates. Narratives recycle. Audiences become increasingly parasocial. Public performance continues while public life gradually disappears.

Zombie publics are highly visible precisely because they possess relatively little interiority. They are optimized for outward radiation rather than inward development.

The third feature is artificial intelligence.

The emergence of large language models has often been interpreted as the culmination of the public internet. It is almost the opposite.

Black holes are not entirely black. Quantum mechanically, they emit Hawking radiation. This radiation does not consist of messages sent from beyond the event horizon. It is the long thermodynamic aftermath of gravitational collapse itself.

Artificial intelligence increasingly occupies an analogous role: That of a thermalized fossil public.

Its training corpus consists overwhelmingly of the accumulated public internet that existed before blackholification reached its present stage: books, websites, Wikipedia, blogs, forums, public code repositories, digitized archives, public conversations, and institutional documents. Models continuously remix this material into fluent statistical syntheses. They possess extraordinary knowledge of the fossil public.

What they fundamentally lack access to is the living interiority of today’s blackholifying cozyweb.

This is not a temporary engineering limitation. It is a consequence of the causal geometry. The defining conversations of mature communities increasingly occur beyond event horizons inaccessible to public observation. AI therefore becomes the thermalization of the fossil public: the ambient informational glow emitted by a civilization whose most vital conversations have already disappeared into causally disconnected interiors.

This also explains why AI often feels strangely omniscient yet oddly lifeless. It has absorbed the archaeological record of public civilization while remaining largely excluded from its present tense.

This distinction also clarifies the status of genuine leaks. Screenshots from private Discords, leaked Slack logs, internal documents, accidental recordings, whistleblower disclosures—these are not Hawking radiation. They are better understood as fragments of the surrounding black-hole system that never fully crossed the horizon: material lingering in unstable orbit within the accretion disk, occasionally perturbed outward before finally disappearing. They are exceptional precisely because they do not violate the causal integrity of the interior, and can be flexibly narrativized in ways entirely disconnected from the interior narrative. The event horizon remains intact.

Taken together, these three phenomena define the observable universe of the Dead Forest.

1. First, the thermalized fossil public continuously recirculated by artificial intelligence.

2. Second, the liminal accretion disks of boundary objects orbiting living communities.

3. Third, the zombie public whose endless performances preserve visibility while generating progressively less shared reality.

What is conspicuously absent is the thing that once defined the internet itself: a common causal manifold in which strangers could reliably become contemporaries through public communication.

The internet has not become private. It is dying with cosmological grandeur.

The public did not disappear because everyone retreated into private spaces. It disappeared because those spaces underwent gravitational collapse into compact worlds whose interior histories increasingly belong only to themselves. We still observe their radiation. We still see the debris orbiting their boundaries. We still mistake the theater of zombie publicity for public life.

But we no longer inhabit a universe in which the public is the primary medium through which reality is jointly constructed and enacted.

The Dead Forest is what remains after the public has collapsed into black holes.

***

The Dead Forest did not emerge because a new force entered history. It emerged because the forces that produced the Dark Forest were allowed to operate uninterrupted until they exhausted the geometry of the public sphere itself.

The original diagnosis remains largely intact.

Search dissolved into recommendation. Recommendation dissolved into algorithmic manipulation. Surveillance capitalism transformed every public utterance into extractable behavioral data. Culture-war dynamics converted visibility into permanent reputational exposure. Institutions lost the capacity to sustain neutral public ground. Politics ceased to be one domain among many and became the organizing logic of nearly every public conversation. Social media steadily rewarded identities optimized for conflict rather than curiosity. The internet of beefs expanded until it ceased to be merely an internet phenomenon and became a general model for social life.

The cozyweb was the rational adaptation.

People withdrew into smaller spaces where trust could once again be accumulated rather than continuously spent. Communities became increasingly bounded, invitation-based, contextual, and difficult to search or index. Public writing increasingly served not as participation in a common discourse but as boundary maintenance, recruitment, diplomacy, fundraising, publishing, or reputation management on behalf of private interiors.

The public sphere was no longer where life happened. It became where communities advertised their existence.

Nothing fundamentally changed after this decade-old diagnosis. No creative response took shape to check it. The dynamics simply continued unchecked as the no-treatment prognosis suggested. Cozy spaces simply accumulated enough cultural matter to undergo gravitational collapse.

COVID accelerated the migration of meaningful relationships into digitally mediated private spaces. Remote work replaced organizational corridors with Slack workspaces. Institutions weakened further while informal affinity networks strengthened. The second Trump era completed the normalization of permanent political mobilization as the background condition of public life. Meanwhile, every advance in generative AI increased the economic value of public text while simultaneously reducing the incentive to produce genuinely new public writing. The public web became both more extractable and less generative.

The result was not a new equilibrium but a phase transition.

Dark Forest Theory described a world in which everyone remained connected but increasingly chose silence. Dead Forest Theory describes the world after enough silence has accumulated that the public itself loses coherence as a shared causal medium.

AI did not produce this transition. It merely paved the dead cowpaths. Large language models arrived only after the living public had already begun collapsing irreversibly into cozy interiors. It industrialized the recycling of the fossil public while accelerating the exhaustion of what remained outside the horizons.

The internet did not die because of AI. AI is inheriting the remains as it dies, and the cycling of archival and carnival time winds down into a terminal archive.

***

If the Dead Forest is the endgame of the internet we inherited, the obvious question is whether another public sphere can ever emerge.

Not whether this public can be repaired. Cosmology suggests it cannot. Black holes do not become stars again. Gall’s law strengthens this intuition:

“A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works and cannot be patched up to make it work. You have to start over with a working simple system.”

The question is whether history can produce another arborescent digital cosmos and whether we can seed a new forest right now.

Nearly every civilization has imagined a cosmic tree: the Norse Yggdrasil, the Indian Kalpataru, the Mayan Ceiba, the Persian Gaokerena, the Biblical Tree of Life. Their details differ, but they share a common structural intuition. The cosmos is not fundamentally a collection of disconnected places. It is one living organism whose branches connect many domains without erasing their differences. The tree is neither a centralized empire nor an archipelago. It is a common living medium.

The public internet briefly approximated such a structure. We mistook it for a permanent feature of technological civilization. It now appears more likely to have been an unusually low-entropy historical accident.

If another cosmic arborescence is to emerge, it will not do so by reversing the gravitational collapse of the present one. It must grow from whatever remains outside the horizons before those remnants themselves disappear.

This suggests less a legible program of obvious actions than a set of six simultaneous grand challenges that require genuine invention to address.

The first concerns media.

The next public medium cannot simply optimize engagement more efficiently than its predecessors. Nor can it merely federate today’s cozywebs. It must possess intrinsic anti-cozy properties: mechanisms that continuously regenerate encounters between strangers without collapsing into algorithmic extraction or culture-war dynamics. Publicity itself must become renewable rather than exhaustible.

The second concerns politics.

The internet cannot recover a public if politics remains organized around permanent mobilization. A society in which every public utterance is interpreted primarily as coalition signaling cannot sustain common causal space. Any successor public must make disagreement productive without making identity existential. It must allow for mutual co-existence in citizenship rather than a condition of endemic armed activism.

The third concerns artificial intelligence.

Today’s models increasingly thermalize the fossil public. A future public intelligence would instead require access to continuously renewed living culture without simply consuming or exposing it. This is not merely a data problem. It is a civilizational design problem. Intelligence must become metabolically coupled to public life rather than archaeologically dependent upon its remains.

The fourth concerns institutions.

Public institutions once served as long-lived repositories of common knowledge whose legitimacy exceeded that of any particular community. Most now either retreat into cozy interiors themselves or perform zombie publicity in order to remain visible. A new arborescence requires institutions capable of producing genuine common reality rather than merely broadcasting legitimacy.

The fifth concerns public life itself.

Zombie publics cannot simply be replaced by better influencers, healthier discourse, or more responsible platforms. The problem is ontological rather than behavioral. Public life must once again become a place where significant interiority can develop rather than merely be represented. People must once again possess reasons to conduct meaningful portions of their intellectual, artistic, scientific, and civic lives in public.

Finally, there is the challenge of time.

Every year, more communities pass beyond their event horizons. More knowledge is born irretrievably private. More institutions become performative. More AI systems are trained on increasingly recycled corpora. More of the accretion disk spirals inward. More of the fossil public becomes thermalized.

The urgency is cosmological.

Dead forestification appears to possess positive feedback loops. Every successful retreat into interiority increases the incentives for further retreat. Every reduction in the vitality of the public increases the relative value of private worlds. Every increment of AI-generated public text reduces the density of genuinely renewable public culture available for future intelligences. Every new black hole slightly alters the geometry through which subsequent ones form.

If there is a threshold beyond which no new cosmic tree can grow, we do not know where it lies.

Nor do we know whether we have already crossed the ultimate event horizon on the trajectory to collective social heat death of the forest we inhabit now.

The task, then, is not to restore the internet we lost. It is to preserve enough living matter outside the horizons that another cosmology remains possible. The myths of Yggdrasil and Kalpataru remind us that civilizations have long imagined worlds held together by living connective tissue rather than by force or by isolation. Whether technological civilization can grow such a tree again is the defining grand challenge of the twenty-first century.

Dead Forest Theory suggests that this forest cannot be saved, but holds out the possibility that a new one can still be planted before it dies.

Many years ago, in an earlier era of silliness about a looming ed-tech revolution, a startup founder sent me an email, asking that I please hold off from writing anything about the company. They'd experienced a huge surge in demand for their app – it was fall, students were headed back-to-school. He confessed that, until their new round of funding actually cleared the bank, they were struggling to pay their bills, particularly a very large bill for Twilio, the messaging service that was, in fact, almost the entirety of the functionality of what they were offering to schools. But unlike Twilio, this particular app was "free" – I put that word in quotation marks because, of course, nothing is. There are real costs to running a business: paychecks have to be cut, operational expenses have to be paid for. What enabled this specific startup, and so many like it, to offer their product at no charge was a subsidy of sorts, in the form of venture capital, which has promoted a new, strange business model, allowing companies to avoid thinking about revenue let alone profitability – ostensibly just in the early stages, but increasingly (and dangerously), long into the lifespan of a company.

Schools love "free," of course, as they're almost always operating with budgetary constraints (and/or with political threats to axe even more of their funding). So "free" software has been able to make substantial in-roads into classrooms -- particularly alongside the narrative about the compulsory usage of ed-tech -- and many educators and administrators have been able to convince themselves that there really are no costs to adoption.

But at some point, the venture capital runs out. At some point, the bill comes due. At some point, a startup has to determine what part of the product or service it will charge for. "Find the thing that schools can't do without," I heard one founder quip at an ed-tech event years ago, "and make them pay for that." "Make it painful for them to stop," he added, suggesting that students with disabilities made for a very good target for this strategy.

Even when these pricing practices aren't nefarious (although frankly they mostly are -- that's capitalism), we all understand how software can be "painful to stop" -- not because we're "addicted" (I will always hate that framwork) or because of some weird "brand loyalty," but simply because we have used a particular tool for so long; we are familiar with its functionality, and we don't want to have to learn a new interface; and, of course, these apps have so much of our work, our data, and tech companies have made it virtually impossible to move any of it elsewhere.

There are other problems with "free" ed-tech too. As it costs nothing to adopt (and then likely reject) different apps, schools have found themselves using thousands of different apps (an average of almost 2600 according to a report a few years ago -- that report found that students accessed about fifty unique applications over the course of the year). Student and teacher data is strewn across these tools, often with little regard for privacy or security, let alone any semblance of curricular consistency or a shared educational experience.

This flurry of tool adoption is typically (mis-)read as enthusiasm for ed-tech rather than what's far more likely: most school software is shit; you download it, use it once or twice, and yikes, you never touch it again. And even when there is enthusiasm -- such seems to be the case with students' usage of ChatGPT to do their homework -- we should still pause before celebrating this as some educational marvel. We should still pause before interpreting any of this as indication there's some insatiable demand for ed-tech or for generative AI because it's good. Mostly, it's because it's free.

As Dave Karpf wrote back in May, "we are still in the free-trial-period of most AI products. Everything is being subsidized. That renders a false picture of the demand curve." "Free" distorts demand; "free" also distorts a business's viability. AI companies -- much like many ed-tech startups who like to boast, come August, an explosion in sign-ups for their free app -- might be able to show off an incredible growth in the number of subscriptions; but that growth is actually driving their revenue down not up.

(And now, as AI companies are starting to charge based on token-usage not simply a flat subscription, many of their customers are shocked at the bill. "Companies are scrambling to stop spending so much on AI," as 404 Media reports.)

What we are witnessing with AI is a demand for compute, the AI companies' demand for compute: their demand for data centers, their demand for data storage, their demand for data processing. Generative AI seems to have an insatiable demand for power (literally and figuratively) and for resources; and the latter isn't just about water -- the industry has also been gobbling up the components necessary to build all computer hardware, driving the prices up for everything from hard-drives to chips. All computer hardware -- from smartphones to laptops to video game consoles, obviously, but also all the appliances and vehicles and gadgets that we have (unnecessarily) stuck computers into so that dangerous men could chortle that "software is eating the world."

If software has done anything like this over the past few decades, it's because hardware -- consumer electronics, broadly speaking -- has become fairly affordable. So affordable that people buy the latest model even if its capabilities are only marginally better than the one they already possess. And importantly (for our purposes here, at least) so affordable that schools have been able to bring ed-tech out of "the computer room" and into (almost) every classroom and onto (almost) every desk.

An increase in the price of new computers – see related reporting this week from Wired's Boone Ashworth, New York Mag's John Herrman, and Disconnect's Paris Marx – indeed, in some cases, the inability to even make new devices, comes at a pretty rough time for ed-tech sales, I reckon, particularly as more and more parents and teachers are prompting more and states and districts to push back on "screen time."

What looks like a bad investment academically and increasingly like a bad investment politically may also soon be calculated as a bad investment financially as well.

People have been told for decades now that we cannot afford not to have students on computers. But it might be, if the AI industry continues maniacally apace, that none of us will be able to be on computers at all.

---

In other news:

Via The Guardian: "Screen time can damage under-twos’ development, landmark study suggests"

Via Wired: "Meta Contractors Posed as Teens to Prompt Rival Chatbots About Suicide, Sex, and Drugs"

"County With 37 Data Centers Asks Schools to ‘Conserve Electricity’," 404 Media reports

"A Charter School Spent $500,000 on AI-Powered Humanoid Robots. Was It Worth It?" asks Voice of San Diego. (Is that a serious question?!)

---

This week's bird is the raso lark, a small bird found only on the Raso islet of the Cape Verde islands. Very little is known about this critically endangered bird – its desert home is remote, and there hasn't been much ornithological work on the archipelago, Wikipedia reports. The bird is threatened by geckos and kestrels, as well as by dogs and cats accidentally introduced to the area. Also in critical danger: Cabo Verde, which plays Argentina a little later this evening. How can you not root for this little lark? How can you not root for Vozinha?

Thanks for reading Second Breakfast. Please consider becoming a paid subscriber. Your financial support is what enables me to do this work.

Open Source AI Gap Map

Current AI is "a global partnership building a public option for AI", founded as a non-profit at the AI Action Summit in Paris in February 2025 and backed by serious capital ($400m already committed).

They launched their Gap Map a couple of days ago - an attempt at indexing the current state of open source AI:

The Gap Map v0.1 details 421 products in depth: 266 software tools and libraries, 85 models, 50 datasets, and 20 hardware projects, produced by 228 organizations. These products are organized into 14 categories across 3 layers of the stack (model components, product / UX, and infrastructure). The remaining 24,400 artifacts constitute the uncategorized long tail of the open source AI ecosystem, and will carry no score until they are researched and cited.

The map itself is interesting to explore, but I'm more excited about the underlying data - released under an MIT license in the currentai-org/os-ai-map GitHub account: 1,184 YAML files plus the notebooks, schemas and other scripts used to help gather them.

Since the files are on GitHub you can use Datasette Lite to explore some of them - here are 16,185 GitHub repos the project is tracking as a CSV file loaded into Datasette Lite.

Tags: open-source, ai, datasette-lite, generative-ai, local-llms, llms