Sharon Machlis

 • LinkedIn
 • GitHub
 • Mastodon
 • Twitter (not currently active)

 

About Sharon Machlis

Sharon is an award-winning journalist and data analyst who is equally at home analyzing data, coding tools for journalists, and covering technology. She is currently director of editorial data & analytics at Foundry (an IDG Inc. company), which publishes tech Web sites including Computerworld, CIO, PCWorld, and Macworld; and author of InfoWorld's Do More With R series .

Her book, Practical R for Mass Communication and Journalism, is available from publisher CRC Press and Amazon (you can see Six chapters free online).

Sharon is well known in the R community. She has taught workshops at ProPublica, and Investigative Reporters and Editors conferences and keynoted the 2020 European R User Meeting.

Sharon received an ASBPE national gold award for impact/investigative online excellence (see story) and two ASBPE national golds for best how-to article (see 2014 and 2017 winners) She was also named the Digital Analytics Association's top practitioner in 2021 for her work analyzing data at her job and in her community.

Sharon holds an Extra-class ham radio license and was honored by the Association of Radio Amateurs of Bosnia & Herzegovina "for extraordinary contribution to transmitting humanitarian messages of the citizens of Bosnia-Herzegovina" during the 1992-95 war.

Sharon's other hobbies include photography, travel, hiking, snowshoeing, crocheting, and classical piano. And, she's somewhat obsessed with both generative AI and the R programming language.

You can follow Sharon on Mastodon at @smach@masto.machlis.com and her searchable app with her Mastodon posts, as well as on LinkedIn. (She is not currently active on Twitter, but you can see her prior tweets at @sharon000.)

Sharon's Recent InfoWorld and Computerworld Articles

5 easy ways to run an LLM locally 2024-03-28

How to run R in Visual Studio Code 2024-02-15

Posit lays off R Markdown, knitr creator Yihui Xie 2024-01-05

8 ChatGPT tools for R programming 2023-12-21

Anthropic's Claude 2.1 LLM turbocharges performance, offers beta tool use 2023-11-21

Anthropic's Claude 2.1 LLM turbocharges performance, offers beta tool use 2023-11-21


Sharon's Data-Related Mastodon Posts

@simon It feels like a “last mile” issue. Open source can make it easier to implement my specific project, but it likely doesn’t accomplish my end goal. There are so many end goals, open source can't handle them all. If my end goal is to create a platform to publish data on the Web, Datasette pre-empted that. But more end goals are last-mile “I want to publish my city's data on the Web.” Datasette makes it easier for a Web dev to do that. An LLM can create the last-mile page. (2024-04-22 12:37:31)  >>

- - - - - - - - - - - - - - - - - -

@gdbassett It's a very interesting question. Some studies suggest that generative AI levels the playing field, helping people at the bottom more than the top; data for other uses seem to show that advanced users benefit. But if you follow @simon , it's hard to argue that great programmers aren't benefitting from genAI. (2024-04-20 20:07:29)  >>

- - - - - - - - - - - - - - - - - -

@jneno Heh. Experiment to see if discussions here can work better than over at That Other Site?

LLMs are not as simple to use as many people first expected, which is a real issue. Is it worth learning to use them well? I think yes. Some others try a prompt or two, are disappointed, and draw conclusions.

Also, anyone who thinks free ChatGPT and GPT-3.5 are the sum total of generative AI will likely have a different outlook than people who have invested time working with frontier LLMs. (2024-04-20 19:56:42)  >>

- - - - - - - - - - - - - - - - - -

@gdbassett Could be. I've always been highly skeptical of blockchain and don’t see the similarities beyond “there's hype”. Dot-com was overhyped too in the ‘90s but still radically transformed retail. Hype ≠ useless.
Did lots of big companies roll out blockchain in a serious way? GitHub Copilot alone has paid users in 37,000+ organizations.
Some early data show gives improved performance in several fields, not just coding. Is that real? Too early to know ultimate impact. (2024-04-20 12:50:26)  >>

- - - - - - - - - - - - - - - - - -

See more posts . . .