Google Colab Promises 'AI-Powered Coding, Free of Charge' (blog.google) 24
Google Colab hosts free cloud-based "executable documents" that, among other things, let you write and run code in your browser (in dozens of languages, including Python).
Over 7 million people, including students, already use Colab, according to a recent post on Google's blog, "and now it's getting even better with advances in AI [with] features like code completions, natural language to code generation and even a code-assisting chatbot."
Google says it will "dramatically increase programming speed, quality, and comprehension." Our first features will focus on code generation. Natural language to code generation helps you generate larger blocks of code, writing whole functions from comments or prompts. [For example: "import data.csv as a dataframe."] The goal here is to reduce the need for writing repetitive code, so you can focus on the more interesting parts of programming and data science. Eligible users in Colab will see a new "Generate" button in their notebooks, allowing them to enter any text prompt to generate code.
For eligible paid users, as you type, you'll see autocomplete suggestions.
We're also bringing the helpfulness of a chatbot directly into Colab. Soon, you'll be able to ask questions directly in Colab like, "How do I import data from Google Sheets?" or "How do I filter a Pandas DataFrame?"
Anyone with an internet connection can access Colab, and use it free of charge... Access to these features will roll out gradually in the coming months, starting with our paid subscribers in the U.S. and then expanding into the free-of-charge tier.
It's powered by Google's "next generation" machine-learning language model PaLM 2 (announced earlier this month), which "excels at popular programming languages like Python and JavaScript, but can also generate specialized code in languages like Prolog, Fortran and Verilog." Colab will use Codey, a family of code models built on PaLM 2... fine-tuned on a large dataset of high quality, permissively licensed code from external sources to improve performance on coding tasks. Plus, the versions of Codey being used to power Colab have been customized especially for Python and for Colab-specific uses.
Over 7 million people, including students, already use Colab, according to a recent post on Google's blog, "and now it's getting even better with advances in AI [with] features like code completions, natural language to code generation and even a code-assisting chatbot."
Google says it will "dramatically increase programming speed, quality, and comprehension." Our first features will focus on code generation. Natural language to code generation helps you generate larger blocks of code, writing whole functions from comments or prompts. [For example: "import data.csv as a dataframe."] The goal here is to reduce the need for writing repetitive code, so you can focus on the more interesting parts of programming and data science. Eligible users in Colab will see a new "Generate" button in their notebooks, allowing them to enter any text prompt to generate code.
For eligible paid users, as you type, you'll see autocomplete suggestions.
We're also bringing the helpfulness of a chatbot directly into Colab. Soon, you'll be able to ask questions directly in Colab like, "How do I import data from Google Sheets?" or "How do I filter a Pandas DataFrame?"
Anyone with an internet connection can access Colab, and use it free of charge... Access to these features will roll out gradually in the coming months, starting with our paid subscribers in the U.S. and then expanding into the free-of-charge tier.
It's powered by Google's "next generation" machine-learning language model PaLM 2 (announced earlier this month), which "excels at popular programming languages like Python and JavaScript, but can also generate specialized code in languages like Prolog, Fortran and Verilog." Colab will use Codey, a family of code models built on PaLM 2... fine-tuned on a large dataset of high quality, permissively licensed code from external sources to improve performance on coding tasks. Plus, the versions of Codey being used to power Colab have been customized especially for Python and for Colab-specific uses.
Sorry Google! (Score:3)
Re: (Score:3)
Same and I have it disabled in Emacs, etc...
For me, typing things out gives me more time to think about what I'm coding -- and not coding.
Re: (Score:2)
Re: (Score:1)
Re: It's FREE!!! (Score:5, Interesting)
Basically Google gets your code, which they will likely use to replace coders/programmers. You generate the code, then fix and send it back for validation. AI gets trained off said code to improve its generative coding abilities, which will no doubt be used to replace coders/programmers. Those that are already established SW developers at Google and the like can meanwhile just use that code for their projects, all without necessarily having to recruit more manpower. License the service to other companies and bam, you have another source of revenue.
In essence, new and coming coders/programmers are effectively training their replacement for free before they can even get the job. The user really is nothing more then a product.
Re: (Score:3)
It shouldnt be lost on *anyone* that most amateur and low-budget AI research is done on Google colab (how else are you gonna get your hands on a H100 GPU farm?).
That said, they DO have a monetization model. You pay for GPU credits when running models. I *think* you can run toy models (A few gigs in size), but thats it. If you want to play with big models like trying to fine tune Llama65G or whatever, thats gonna cost.
They get your code/needs/etc, you get ... (Score:4, Insightful)
Re: (Score:2)
Yeah, but it's totally legal and legacy corporate espionage is not.
Uh huh. (Score:3)
Google says it will "dramatically increase programming speed, quality, and comprehension."
Betting it will still be constrained by the triangle: fast, good, or cheap -- pick two.
Re: (Score:2)
FUD (Score:3)
Re: (Score:3)
They are bullshitting people. The amount of bullshit surrounding generative AI surpasses even the bullshit generated by the generative AI. (Bullshit being defined as content whose producer is not necessarily deceptive, he just doesn't care if he says is true or false.)
Re: (Score:2)
The amount of bullshit surrounding generative AI surpasses even the bullshit generated by the generative AI
Good one.
You get what you pay for (Score:3)
This is classic "you get what you pay for". Simple boilerplate code, like their "import a CSV into a dataframe", that's easy. It's a well-known problem with a well-known solution, all the developer needs to do is fill in the blanks for column/field names and data types. Getting it to generate code to transfer field values from one form into another, where the field names are user-entered values, the number and order of fields is different between the two, field names may be duplicated and there's no requirement that there be the same number of fields with a given name in both forms, and there are no well-defined rules for determining whether two field names are similar enough to match, THAT is more of a challenge. Call me when their AI can solve that problem reliably to the user's satisfaction.
Re: (Score:2)
The other issue of course is that for any program beyond basic or boilerplate code the human language description will be vastly longer and more prone to ambiguities than the actual code itself so I'm not really sure how much use this will be anyway. As you say, boilerplate generators already exist so wheres the gain for the user here?
Methinks they promise too much (Score:2)
"dramatically increase programming speed, quality, and comprehension" sounds like a blatant marketing lie, nothing else. Better stay away from this, because clearly you are the product and they have zero accountability if the fuck you up.
Re: (Score:1)
Shows they are not just bullshitters but charlatans, charlatans being defined as those who promise only the good without any of the bad.
Re: (Score:2)
Indeed. That seems a very adequate term for this.
Re: (Score:2)
By the time you get used to it (Score:2)
you'll get the notification that the project will be abandoned within the next 6 months.
Google gets me everytime with the commitment to their own products.
Hey, Colab. (Score:2)
Write a version of your code that's 5% as capable in every way,
then replace yourself with the new, improved version,
then repeat this request.
Hey Colab, how do I write a back door? (Score:2)
Hey Colab, example code for stealing password?
What could possibly go wrong...
Colab is a Jupyter Notebook with bugs (Score:2)
I instantly reached the limitations of the free resources and ran into multiple code breaking issues with no workarounds, since you can't upgrade or change the versions of the libraries without causing even more problems. Just run your own local env - it's free and you have full control.