According to the developers, Codex can now significantly simplify the work of experienced programmers, as well as help to train beginners. Artificial Intelligence literally turns a stream of words into real software like websites or rudimentary games. For example, a user can describe the basic appearance of a site or its functionality by listing requirements like the location of menus or text boxes using plain spoken English, and Codex will generate a finished product based on its own “insights” into what needs to be accomplished.

In this variant, programming skills are not required at all in some cases. However, basically Codex will only work as an “assistant” or “deputy” programmer, implementing the conceptual ideas of the “creators” in the simplest way possible.

According to Greg Brockman, the technical director and co-founder of OpenAI, the new tool is intended for a wide range of programmers. First of all, it will allow them to get rid of monotonous repetitive work using libraries, functions and APIs.

Codex is based on the GPT-3 algorithm – created by OpenAI for generating “meaningful” texts based on analyzing many terabytes of text material on the Internet. However, it can provoke a negative reaction from developers who put free and open source software online. The algorithm uses additional data that was literally “shoveled” from free software repositories for subsequent analysis.

OpenAI has said it does not violate any copyright or related rights laws. Nevertheless, many people expect criticism from volunteers who created countless lines of code – the company intends to cash in on the efforts of the community, who posted their software to be parsed and used for free. So far, the programmers do not have a common or at least dominant position on the newly created tool.

The post OpenAI has taught AI to convert commands in English into program code appeared first on DM Onzon.

Among the exotic programming languages for writing malware are Go (Golang), D (DLang), Nim and Rust. As a rule, these languages are used to create malicious software aimed at large organizations and which exploits weaknesses in the protection of the corporate infrastructure. We are talking about malware that implements multi-stage, carefully planned and organized cyber-attacks.

Once on the victim’s computer, these software modules bypass traditional defenses to later download, decode and launch remote access Trojans, most often Remcos and NanoCore, as well as Cobalt Strike beacons.

In some cases, when the cybercriminals have sufficient resources, non-standard languages for these purposes are used not only for the modules performing the initial attack, but also for the whole malware package. The Go language is of particular interest to attackers. It is used by hackers on an almost regular basis for all major platforms.

According to the authors of the study, new or unusual programming languages make it difficult to decompile malware, circumvent defenses that use signature-based detection tools with relative ease, and simplify cross-platform compatibility of attacks. The code base of such malware itself is an additional level of disguise. Their developers, in fact, do not have to make any additional effort; the effectiveness of the attack is increased simply by the fact that one of these programming languages is used.

The post Malware developers are increasingly using exotic programming languages appeared first on DM Onzon.