Steven Johnson has a status as a analysis software program nerd. The writer of 13 nonfiction books, he’s consistently searching for digital instruments to streamline his inventive course of. So when massive language fashions – which energy text-generating AI instruments reminiscent of ChatGPT – began getting consideration, he was most desirous about what they might imply for organising info.
In 2022, an article Johnson wrote about LLMs for the New York Occasions caught the attention of researchers at Google Labs, the tech firm’s experimental AI arm, who got here to him with a proposition: would he assist them develop the sort of digital analysis assistant he’d been dreaming of? The result’s NotebookLM, a note-taking instrument that makes use of AI to assist organise, summarise and reply questions on any info you give it. “The way in which we give it some thought is it’s a instrument for understanding issues,” Johnson says.
Generative AI has been eagerly adopted within the productivity tech house, with new and current instruments promising options that can assist streamline your schedule, simplify emails and take notes extra successfully. I’ve typically been sceptical of such merchandise, assuming that by the point I’ve fiddled round to get them arrange, I’ve in all probability negated most of their productivity advantages. However with a brand new era on the scene, I puzzled if my workflow couldn’t profit from an AI increase. Might these instruments truly ship a number of the benefits claimed by AI acolytes, automating extra humdrum points of labor and leaving me free to deal with inventive endeavours?
I approached NotebookLM with warning, noting the disclaimer on the backside of the display: “NotebookLM should typically give inaccurate responses, so it’s your decision to verify any details independently.” As a journalist, I’m cautious of AI’s tendency to “hallucinate”. However NotebookLM differs from textual content turbines reminiscent of ChatGPT, Claude or Gemini in that it solely works with the data you present. You can begin a brand new “pocket book” for a mission and add up to 50 “sources”. These can embody pdfs, audio recordsdata (which it routinely transcribes), .txt recordsdata, internet hyperlinks and Google Docs. I attempted importing notes and interview recordsdata for a narrative I used to be engaged on, then requested the “Pocket book information” – an AI assistant – to produce a briefing doc primarily based on their content material.
I instantly noticed the enchantment. The temporary summarised key factors raised from hours of audio and drew hyperlinks between my sources. It even pulled out key quotes. Utilizing the AI chat operate, I interrogated the fabric additional, asking which findings have been most stunning and posing particular questions in regards to the content material. Reasonably than trawling by my transcripts to bear in mind which interviewees had made a specific level, I merely requested: “Who spoke about [X topic]?” and bought an in depth abstract of who stated what – together with footnotes directing me to the precise wording within the authentic materials.
It’s the restriction on the core of NotebookLM – utilizing solely the data you present – that makes it so helpful. I might see it being adopted by college students and information staff of all stripes. Johnson says he has a pocket book for his subsequent e book, the place he shops concepts and asks the information to counsel chapter buildings or key characters. “That’s sort of like a focused brainstorming pocket book,” he says. He additionally has one for the NotebookLM mission itself, the place he collates bulletins and memos. He used the information to draft a press launch for a brand new function, and bought it to give you its personal tagline: “Suppose smarter not tougher.”
There are drawbacks. The information provides lengthy solutions to queries, which ends up in a variety of textual content to wade by, and I discovered a degree of subjectivity creeping into the AI-generated summaries (it deemed my interview with Johnson “fascinating”). This was significantly evident within the “Audio Overview” possibility, a much-hyped function that creates a 10-minute AI-generated podcast primarily based on your sources. It’s pretty mind-blowing at first: it actually appears like two podcast hosts knowledgably discussing the subject material. However I discovered that it extrapolated from my materials, making worth judgments that didn’t essentially mirror the unique content material.
NotebookLM positively saved me time finding info, however productivity isn’t just about effectivity. Johnson believes the instrument can help creativity on a deeper degree. He has one pocket book that he calls his “all the things pocket book”. It accommodates 1000’s of quotations he’s collected from ebooks over the previous 25 years, and he’s slowly including the textual content of all of the books he’s written. He then treats it like an extension of his reminiscence, utilizing the AI chat to resurface subjects of curiosity or draw connections between sources. “It’s extremely good at simply being like: ‘Oh yeah, 15 years in the past you learn that article about ant colonies,’” he says.
This thought of outsourcing or increasing mind energy is a standard theme in productivity circles. The metaphor of a “second mind” comes up so much. “The idea of the second mind is successfully a manner that allows you to have one other product that does a variety of the heavy lifting for you, liberating you up for extra high-value work,” says Robbie O’Connor at San Francisco-based productivity platform Notion. O’Connor used to work on Google Maps and compares the second mind thought to how that app supercharged our navigational capabilities. “It gave you much more energy, much more freedom, much more reliability,” he says.
Notion provides a set of organisational instruments, with numerous templates for note-taking together with a to-do record, mission planner and finances, in addition to AI options. It additionally has a calendar app and final month introduced an upcoming e-mail app. I discovered it overwhelming after I logged in – I normally simply write a to-do record on paper. As I gamely began filling out a weekly to-do record, behavior tracker and mission planner, I couldn’t assist however really feel I used to be procrastinating reasonably than truly doing the work I used to be planning. A giant a part of Notion’s enchantment, nonetheless, is its collaborative options when adopted throughout groups; working solo, I missed out on these advantages.
after publication promotion
One other note-taking product, Capacities, is extra intuitive, if nonetheless considerably intimidating. This launched in 2022 and not too long ago bought a lift after being recommended by Casey Newton, a tech journalist and productivity instrument power-user. Capacities’s huge thought is to put off the standard file-and-folder laptop storage system. As an alternative, notes are categorised by kind as “objects” – as an illustration, a e book, pdf, individual, weblink, picture or doc. You can then organise and hyperlink these objects by collating them or including tags. The thought is that you need to be ready to rediscover related notes extra simply; they get extra organised over time, reasonably than having a single outlined location from the beginning. “There’s no level in storing and organising stuff if it doesn’t aid you ultimately,” says co-founder Steffen Bleher. There’s additionally an AI assistant.
I began amassing objects, noting down story concepts and dropping in fascinating hyperlinks. It was a great way to maintain issues in a single place; normally I’d jot concepts within the Apple Notes app and bookmark hyperlinks in my browser. However after utilizing Capacities for some time, I discovered I didn’t have that many notes to profit from its information-networking strategy. It was clear it might require long-term dedication to get probably the most out of it. For Bleher, Capacities is mainly a life-style selection; he makes use of it for each skilled and private functions, including ideas about books he’s studying and concepts that come out of discussions with associates. A video tour by co-founder Michael von Hohnhorst exhibits him jotting down reflections on a podcast he has simply listened to and saving favorite quotes and recipes. I can’t think about documenting my life to this diploma. A Capacities blogpost states: “Constructing a second mind takes self-discipline.” Maybe that is what I lack.
Bleher acknowledges {that a} small cohort of productivity power-users are dwarfed by a for much longer tail of people that simply desire a answer to a selected workflow drawback they’re having. Though he clearly falls within the first camp, he takes a practical view to adopting new tech. “Many individuals fall on this productivity entice that they want to have loopy setups and so they want to do extra stuff, and ultimately they mainly procrastinate after which they don’t get something completed,” he says.
Anna Gurun, an affiliate director at HSM Advisory, a future-of-work advisory group, echoes this sentiment. “The massive problem is complicated productivity with busyness,” she says. You would possibly tick off numerous duties in a to-do record app however are you truly reaching what you want to? Your AI calendar would possibly organise your schedule however are you continue to spending an excessive amount of time in pointless conferences? “Should you’re not clear within the outcomes you need to attain then you definately’re not going to have the option to use the instrument successfully,” she says. Productivity additionally has a extra emotional aspect, she provides, which tech instruments can’t actually assist with. Perhaps you’re much less productive since you’re sleeping poorly or you’ve gotten simply been by a nasty breakup. The apps can additionally make failure appear greater, she says: “You kind of enlarge behaviour.”
These phrases got here to thoughts as I seemed again at a calendar instrument I’d been making an attempt out. There are various sensible scheduling instruments on the market – I opted for Reclaim.ai, owned by Dropbox. It syncs with Google Calendar and, when utilized by groups, optimises the perfect assembly instances to swimsuit all attenders. It additionally has a number of easy however wise options, together with the power to routinely add a “decompress” buffer after conferences that embody a video hyperlink and the choice to enter “habits” – common duties you need to reserve time for. I set a day by day behavior to spend an hour engaged on that novel I maintain which means to write, which Reclaim.ai dynamically discovered time for round my different obligations. At first, having the calendar reminder was a very good motivator, however I quickly realised I’d been far too formidable. Nevertheless refined the options, no scheduling instrument can create extra hours within the day, or drive me to truly do the work.
As AI will get smarter, although, it could have the option to deal with extra of these irksome little duties that distract from what we actually need to spend time on. The following pattern is for AI “brokers” – a step on from a chatbot, with the power to truly take motion. This month, AI firm Anthropic demoed an experimental AI model that can routinely fill in an internet kind by discovering knowledge from throughout somebody’s laptop. Google is reportedly creating an AI agent that would autonomously e book a flight or make a web based buy.
Rahul Vohra, founder and chief govt of e-mail app Superhuman, says this strategy will come to productivity instruments. Superhuman already has AI options to assist write, summarise and reply to e-mail, and not too long ago added one which routinely makes calendar occasions primarily based on emailed particulars. AI brokers might do extra. “You can think about an AI agent that triages on your behalf, that schedules on your behalf, that even writes and at some point sends totally written emails on your behalf,” he says.
Need to arrange a gathering or negotiate a contract? Have your bot speak to my bot.