During my postdocs after graduate school, while training undergrad and grad students, I found myself feeling somewhat frustrated at how things were expected to work FAST AND EASY in terms of experiments and other research outcomes, so I became one of those telling younger people how things were before.... with "before" being just 5-10 earlier!
Today, you can do a "pubmed" or even Google search and find papers you want to read or get them emailed to you, and all you have to do is download them and then read on your computer screen or print out (although this availability depends on the journal and whether your institution has purchased subscription for accessing the PDF files). I do like the option of not printing or photocopying to protect the trees, and do most of my reading and writing work on the computer. When I was in grad school however, we had to go to a "library" (remember those?) to get the journal issue in printed form, make photocopies and then read. The overall time necessary was much longer, and also much better for your health as it involved walking to and from the library, photocopy machine etc.
Everybody had a lab notebook to write the details of each day's work (experiments, pictures of results of gels or bacteria plates taken in a photo room and then pasted onto the notebook) - now it's all mostly recorded in the computer, photos are taken with a digital camera or a fancy scanner and it all goes into electronic files, which are also necessary for publishing. It's convenient, fast and easy but more automated, but I miss the feeling of individuality we got when reading someone's handwriting in a lab protocol or experiment notes. Sending manuscripts to peer-reviewed journals for publication (or grant application for research funding) is now done mostly online, which is faster and once again saves a lot of trees.
Seminars and presentations were more personal, people made drawings and diagrams by hand or took pictures that were "developed" in a dark room and then made slides to use with a projector. Now it's all powerpoint.
In terms of techniques such as the basic ones to "extract DNA", when I first started working in a molecular biology laboratory there was a long "protocol". We started by making a bunch of solutions and buffers to use for the experiment: home (lab) -made solutions. We learned how to make them, what was in them and also what each of the steps meant and did to the cells that we needed to break to release the DNA and purify it away from other intracellular elements such as organelles, proteins and RNA. Other techniques we used to do ourselves included "DNA sequencing". As time went by, biotech companies started making "kits" which come in a box with solutions 1,2,3 and a protocol you follow that tells you how much to add, what to do (mix, centrifuge, incubate for x min, etc). They rarely tell you what's in the solutions or resins (in fact, that's the whole point) they are expensive and young scientists are not aware what each step of the procedure actually does. Now most labs, if money is not a problem, use kits and lab staff demands to stock on kits. On the other hand, it makes results more standardized and comparable. Procedures that took days now take hours, and again that is a good thing but it results in more automated procedures. There are many "robots" available that would process tens or hundreds of samples simultaneously in less than a day. These are very useful for "high through-put screening" used frequently for drug discovery. Once an assay is setup for activity of an enzyme you are interested in, you can "screen" for an inhibitor to use in patients in the future. Literally millions of compounds can be assessed in multi-well plates in these "in vitro" assays.
It used to take months to sequence ONE GENE. Now WHOLE ORGANISMS are sequenced faster and faster, and the DNA sequences are entered into databases that are often accessible to investigators. Next step: individuals of the same species (like us! men and women) can now have their WHOLE genome sequenced in a relatively short time, an ethically controversial topic in terms of how much information you might want (or not want!) to know regarding predisposition/risk to certain diseases.
The amount of data one needed to publish papers was smaller then and much bigger now, as things can be done faster and everything is more competitive. There is usually the need to have a more "complete" story, which has resulted in more pressure for publications to include different fancy techniques and experiments confirming the story presented. Individual labs now either pay facilities to do analysis for them from a submitted sample (cells, cultures, purified RNA or DNA, tissue samples, etc) and/or collaborate with other labs to generate more results and then publish the work (a bigger story) all together. There is also more interaction with mathematical modeling teams in terms of analyzing the results or predicting them based on DNA sequences for example, or protein shapes and structures.
And of course the way things are now will no longer be standard in 5-10 more years...