AWS today announced that CodeGuru, a set of tools that use machine learning to automatically review code for bugs and suggest potential optimizations, is now generally available. The tool launched into preview at AWS re:Invent last December.

CodeGuru consists of two tools, Reviewer and Profiler, and those names pretty much describe exactly what they do. To build Reviewer, the AWS team actually trained its algorithm with the help of code from more than 10,000 open source projects on GitHub, as well as reviews from Amazon’s own internal codebase.

“Even for a large organization like Amazon, it’s challenging to have enough experienced developers with enough free time to do code reviews, given the amount of code that gets written every day,” the company notes in today’s announcement. “And even the most experienced reviewers miss problems before they impact customer-facing applications, resulting in bugs and performance issues.”

To use CodeGuru, developers continue to commit their code to their repository of choice, no matter whether that’s GitHub, Bitbucket Cloud, AWS’s own CodeCommit or another service. CodeGuru Reviewer then analyzes that code, tries to find bugs and, if it does, it will also offer potential fixes. All of this is done within the context of the code repository, so CodeGuru will create a GitHub pull request, for example, and add a comment to that pull request with some more info about the bug and potential fixes.

To train the machine learning model, users can also provide CodeGuru with some basic feedback, though we’re mostly talking “thumbs up” and “thumbs down” here.

The CodeGuru Application Profiler has a somewhat different mission. It is meant to help developers figure out where there might be some inefficiencies in their code and identify the most expensive lines of code. This includes support for serverless platforms like AWS Lambda and Fargate.

One feature the team added since it first announced CodeGuru is that Profiler now attaches an estimated dollar amount to the lines of unoptimized code.

“Our customers develop and run a lot of applications that include millions and millions of lines of code. Ensuring the quality and efficiency of that code is incredibly important, as bugs and inefficiencies in even a few lines of code can be very costly. Today, the methods for identifying code quality issues are time-consuming, manual, and error-prone, especially at scale,” said Swami Sivasubramanian, vice president, Amazon Machine Learning, in today’s announcement. “CodeGuru combines Amazon’s decades of experience developing and deploying applications at scale with considerable machine learning expertise to give customers a service that improves software quality, delights their customers with better application performance, and eliminates their most expensive lines of code.”

AWS says a number of companies started using CodeGuru during the preview period. These include the likes of Atlassian, EagleDream and DevFactory.

“While code reviews from our development team do a great job of preventing bugs from reaching production, it’s not always possible to predict how systems will behave under stress or manage complex data shapes, especially as we have multiple deployments per day,” said Zak Islam, head of Engineering, Tech Teams, at Atlassian. “When we detect anomalies in production, we have been able to reduce the investigation time from days to hours and sometimes minutes thanks to Amazon CodeGuru’s continuous profiling feature. Our developers now focus more of their energy on delivering differentiated capabilities and less time investigating problems in our production environment.”

Image Credits: AWS

A big problem for companies these days is finding ways to connect to various data sources to their data repositories, and Fivetran is a startup with a solution to solve that very problem. No surprise then that even during a pandemic, the company announced today that it has raised $100 million Series C on a $1.2 billion valuation.

The company didn’t mess around with top flight firms Andreessen Horowitz and General Catalyst leading the investment with participation from existing investors CEAS Investments and Matrix Partners. Today’s money brings the total raised so far to $163 million, according to the company.

Martin Cassado from a16z described the company succinctly in a blog post he wrote after its $44 million Series B in September 2019, which his firm also participated in. “Fivetran is a SaaS service that connects to the critical data sources in an organization, pulls and processes all the data, and then dumps it into a warehouse (e.g., Snowflake, BigQuery or RedShift) for SQL access and further transformations, if needed. If data is the new oil, then Fivetran is the pipes that get it from the source to the refinery,” he wrote.

Writing in a blog post today announcing the new funding, CEO George Fraser added that in spite of current conditions, the company has continued to add customers. “Despite recent economic uncertainty, Fivetran has continued to grow rapidly as customers see the opportunity to reduce their total cost of ownership by adopting our product in place of highly customized, in-house ETL pipelines that require constant maintenance,” he wrote.

In fact, the company reports 75% customer growth over the prior 12 months. It now has over 1100 customers, which is a pretty good benchmark for a Series C company. Customers include Databricks, DocuSign, Forever 21, Square, Udacity and Urban Outfitters, crossing a variety of verticals.

Fivetran hopes to continue to build new data connectors as it expands the reach of its product and to push into new markets, even in the midst of today’s economic climate. With $100 million in the bank, it should have enough runway to ride this out, while expanding where it makes sense.

When Apple confirmed it had acquired Fleetsmith, a mobile device management vendor, on Wednesday, it seemed like a straightforward purchase, but Fleetsmith customers quickly learned a key piece of functionality had stopped working  — and many weren’t happy about it.

Apple systems administrators began complaining on social media on the morning of the acquisition announcement that the company was no longer allowing them to connect to third-party applications.

“Primarily Fleetsmith maintained a third-party app catalog, so you could deploy things like Chrome or Zoom to your Macs, and Fleetsmith would maintain security updates for those apps. This was the main reason we purchased Fleetsmith,” a Fleetsmith customer told TechCrunch.

The customer added that the company described this functionality as a major feature in a company blog post:

For apps like Chrome, which are managed through the Fleetsmith Catalog, we handle all aspects of testing, packaging, triage, and deployment automatically. Whenever there’s an update (including security patches), we quickly add them to the Catalog so that our customers can enforce the latest version. In this case, we had the Chrome 78.0.3904.87 patch up within a couple hours of the update dropping.

As one system administrator pointed out, being able to manage Chrome browser security in an automated way was a huge part of this, and that was also removed along with third party app support.

As it turned out, Apple had made it clear that it was discontinuing this feature in an email to Fleetsmith customers on the day of the transition. The email included links to several help articles that were supposed to assist admins with the transition. (The email is included in full at the end of this article.)

The general consensus among admins that I spoke to was that these articles were not terribly helpful. While they described a way to fix the issues, they said that Apple has turned what was a highly automated experience into a highly manual one, effectively eliminating the speed and ease of use advantage of having the update feature in the first place.

Apple did confirm that it had responded to some help ticket requests after the changes this week, saying that it would soon restore some configurations for Catalog apps, and was working with impacted customers as needed. The company did not make clear, however, why they removed this functionality in the first place.

Fleetsmith offered a couple of key features that appealed to Mac system administrators. For starters, it let them set up new Macs automatically out of the box. This allows them to ship a new Mac or other Apple device, and as soon as the employee powers it up and connects to Wi-Fi, it connects to Fleetsmith, where systems administrators can track usage and updates. In addition, it allowed System Administrators to enforce Apple security and OS updates on company devices.

What’s more, it could also do the same thing with third-party applications like Google Chrome, Zoom or many others. When these companies pushed a new update, system administrators could make sure all users had the most recent version running on their machines. This is the key functionality that was removed this week.

It’s not clear why Apple chose to strip out these features outlined in the email to customers, but it seems likely that most of this functionality isn’t coming back, other than restoring some configurations for Catalog apps.

Email that went out to Fleetsmith customers the day of the acquisition outlining the changes:

Attempts to reach Fleetsmith founders for comment were unsuccessful. Should that change we will update the article.

Cape Privacy emerged from stealth today after spending two years building a platform for data scientists to privately share encrypted data. The startup also announced $2.95 million in new funding and $2.11 million in funding it got when the business launched in 2018, for a total of $5.06 million raised.

Boldstart Ventures and Version One led the round, with participation from Haystack, Radical Ventures and Faktory Ventures.

Company CEO Ché Wijesinghe says that data science teams often have to deal with data sets that contain sensitive data and share data internally or externally for collaboration purposes. It creates a legal and regulatory data privacy conundrum that Cape Privacy is trying to solve.

“Cape Privacy is a collaboration platform designed to help focus on data privacy for data scientists. So the biggest challenge that people have today from a business perspective is managing privacy policies for machine learning and data science,” Wijesinghe told TechCrunch.

The product breaks down that problem into a couple of key areas. First of all it can take language from lawyers and compliance teams and convert that into code that automatically generates policies about who can see the different types of data in a given data set. What’s more, it has machine learning underpinnings so it also learns about company rules and preferences over time.

It also has a cryptographic privacy component. By wrapping the data with a cryptographic cypher, it lets teams share sensitive data in a safe way without exposing the data to people who shouldn’t be seeing it because of legal or regulatory compliance reasons.

“You can send something to a competitor as an example that’s encrypted, and they’re able to process that encrypted data without decrypting it, so they can train their model on encrypted data,” company co-founder and CTO Gavin Uhma explained.

The company closed the new round in April, which means they were raising in the middle of a pandemic, but it didn’t hurt that they had built the product already and were ready to go to market, and that Uhma and his co-founders had already built a successful startup, GoInstant, which was acquired by Salesforce in 2012. (It’s worth noting that GoInstant debuted at TechCrunch Disrupt in 2011.)

Uhma and his team brought Wijesinghe on board to build the sales and marketing team because, as a technical team, they wanted someone with go-to-market experience running the company so they could concentrate on building product.

The company has 14 employees and is already an all-remote team, so the team didn’t have to adjust at all when the pandemic hit. While it plans to keep hiring fairly limited for the foreseeable future, the company has had a diversity and inclusion plan from the start.

“You have to be intentional about about seeking diversity, so it’s something that when we sit down and map out our hiring and work with recruiters in terms of our pipeline, we really make sure that diversity is one of our objectives. You just have it as a goal, as part of your culture, and it’s something that when we see the picture of the team, we want to see diversity,” he said.

Wijesinghe adds, “As a person of color myself, I’m very sensitive to making sure that we have a very diverse team, not just from a color perspective, but a gender perspective as well.”

The company is gearing up to sell the product  and has paid pilots starting in the coming weeks.

AWS today announced the beta launch of Amazon Honeycode, a new, fully managed low-code/no-code development tool that aims to make it easy for anybody in a company to build their own applications. All of this, of course, is backed by a database in AWS and a web-based, drag-and-drop interface builder.

Developers can build applications for up to 20 users for free. After that, they pay per user and for the storage their applications take up.

Image Credits: Amazon/AWS

“Customers have told us that the need for custom applications far outstrips the capacity of developers to create them,” said AWS VP Larry Augustin in the announcement. “Now with Amazon Honeycode, almost anyone can create powerful custom mobile and web applications without the need to write code.”

Like similar tools, Honeycode provides users with a set of templates for common use cases like to-do list applications, customer trackers, surveys, schedules and inventory management. Traditionally, AWS argues, a lot of businesses have relied on shared spreadsheets to do these things.

“Customers try to solve for the static nature of spreadsheets by emailing them back and forth, but all of the emailing just compounds the inefficiency because email is slow, doesn’t scale, and introduces versioning and data syncing errors,” the company notes in today’s announcement. “As a result, people often prefer having custom applications built, but the demand for custom programming often outstrips developer capacity, creating a situation where teams either need to wait for developers to free up or have to hire expensive consultants to build applications.”

It’s no surprise then that Honeycode uses a spreadsheet view as its core data interface, which makes sense, given how familiar virtually every potential user is with this concept. To manipulate data, users can work with standard spreadsheet-style formulas, which seems to be about the closest the service gets to actual programming. ‘Builders,” as AWS calls Honeycode users, can also set up notifications, reminders and approval workflows within the service.

AWS says these databases can easily scale up to 100,000 rows per workbook. With this, AWS argues, users can then focus on building their applications without having to worry about the underlying infrastructure.

As of now, it doesn’t look like users will be able to bring in any outside data sources, though that may still be on the company’s roadmap. On the other hand, these kinds of integrations would also complicate the process of building an app and it looks like AWS is trying to keep things simple for now.

Honeycode currently only runs in the AWS US West region in Oregon but is coming to other regions soon.

Among Honeycode’s first customers are SmugMug and Slack.

“We’re excited about the opportunity that Amazon Honeycode creates for teams to build apps to drive and adapt to today’s ever-changing business landscape,” said Brad Armstrong, VP of Business and Corporate Development at Slack in today’s release. “We see Amazon Honeycode as a great complement and extension to Slack and are excited about the opportunity to work together to create ways for our joint customers to work more efficiently and to do more with their data than ever before.”