Brief IA
›

Crosspoint Reader: AI Increases Contributions to Sort

Models & LLM·Tom Levy·

Crosspoint Reader: AI Increases Contributions to Sort

Crosspoint Reader: AI Increases Contributions to Sort
Key Takeaways
1Crosspoint Reader surpasses 300,000 installations on small e-readers.
2According to Justin Mitchell, contributions facilitated by language models are burdensome for volunteers.
3The project lists 11 compatible devices equipped with an ESP32-C3 chip.
Why it matters : managers must reject certain technically viable additions because they do not fit the usual use of the devices.

Crosspoint Reader receives code proposals facilitated by language models, with a case of 10,000 submissions in one month reported by its manager, Justin Mitchell. This open-source firmware for small e-readers has surpassed 300,000 installations. Its volunteers must select the additions, including when they are technically feasible.

Volunteers Must Also Reject Code

Justin Mitchell explains that accessible and inexpensive language models now allow users without technical experience to attempt to add their own features to Crosspoint Reader. Several code proposals are submitted each day; many could work in isolation, but it is not possible to integrate all of them. Some automations are so advanced that contributions are sent without the author being aware, according to Mitchell. He cites the case of a contributor who submitted 10,000 pull requests in a month.

Managing these requests and contributions puts significant pressure on the volunteers. Mitchell describes an audience of nearly 500,000 non-technical individuals, many of whom are discovering open-source software through Crosspoint Reader and sometimes expect assistance. Even feasible requests cover a very wide range.

Sorting through the additions is not limited to their technical feasibility. Mitchell mentions well-intentioned and technically viable proposals that are too specific for the common use of an e-reader with fewer than 60 books. According to him, the project is already receiving more code than it can handle: the scarce resources are now "judgment," "trust," and "responsibility."

Devices Impose Very Limited Memory

The compatibility catalog published by Crosspoint Reader on GitHub lists 11 devices, all equipped with an ESP32-C3 system-on-chip. This family of chips, more powerful than the previous ESP32 microcontrollers, features a single RISC-V core operating at a few hundred MHz. The RAM reaches about 400 kB, with approximately 380 kB available. These devices are too limited to run Linux, which explains why Crosspoint Reader does not depend on it.

La lettre du soirGratuit · 19 h

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Among the compact e-readers on the market, the Xteink X4 Classic is sold for $99 or £75, featuring a 4-inch (10.9 cm) screen. The Xteink X3 adopts a credit card format, with a 3.7-inch (9.4 cm) screen. The original systems of many models impose significant restrictions. However, according to Mitchell, many functions requested by users also exceed what the hardware can accommodate.

Software Boosts Sales of Small E-Readers

In recent years, compact e-readers have decreased in size and price. In 2024, the Onyx Palma offered a smartphone-like format for about $300 or £225. Strong sales were followed by the arrival of other models, such as the Boox Picco.

Some credit card-sized devices incorporate magnets to attach to the back of an iPhone or a phone compatible with a magnetic charger mount. Mitchell describes the principle as turning your phone around to read a book instead of checking negative news, referring to it as the "TikTok-ification of reading." He presents these e-readers as fun gadgets in a way that has become rare.

The availability of free, open-source firmware compatible with multiple models contributes to sales by removing many restrictions. Crosspoint Reader plays this role for most of these devices, with some even shipped with a pre-installed version. For others, replacing the software has become an implicitly associated step with the purchase.

Success Forces the Project to Become a Product

At the beginning of the second day of the Ubuntu Summit, Justin Mitchell, manager of Crosspoint Reader, presented the situation during a talk titled "EPUBs, ESP32s, and Everyone is a Developer Now." He estimates the number of users to be nearly 500,000. The forks of the project on GitHub exceeded 1,600 at the start of his talk and then reached 2,300, a number that continues to rise.

This adoption has led the project to evolve from simple open-source code to a product. One example is the Brick Club List, which indicates compatibility between versions and devices: initially an Excel spreadsheet, it has become a website.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.