Keeping the internet open is crucial, and part of being open means Reddit content needs to be accessible to those fostering human learning and researching ways to build community, belonging, and empowerment online. Reddit is a uniquely large and vibrant community that has long been an important space for conversation on the internet. Additionally, using LLMs, ML, and AI allow Reddit to improve the user experience for everyone.
In line with this, Reddit and OpenAI today announced a partnership to benefit both the Reddit and OpenAI user communities…
Perhaps some members of the Reddit user community may not feel the benefits when OpenAI is training on their data.
While people who joined Reddit presumably understood that anyone could view their data, they never imagined that a third party would then process its data for its own purposes.
Oh, but wait a minute. Reddit clarifies things:
This partnership…does not change Reddit’s Data API Terms or Developer Terms, which state content accessed through Reddit’s Data API cannot be used for commercial purposes without Reddit’s approval. API access remains free for non-commercial usage under our published threshold.
If you’re a biometric product marketing expert, or even if you’re not, you’re presumably analyzing the possible effects to your identity/biometric product from the proposed changes to the Biometric Information Privacy Act (BIPA).
As of May 16, the Illinois General Assembly (House and Senate) passed a bill (SB2979) to amend BIPA. It awaits the Governor’s signature.
What is the amendment? Other than defining an “electronic signature,” the main purpose of the bill is to limit damages under BIPA. The new text regarding the “Right of action” codifies the concept of a “single violation.”
(T)he amended law DOES NOT CHANGE “Private Right of Action” so BIPA LIVES!
Companies who violate the strict requirements of BIPA aren’t off the hook. It’s just that the trial lawyers—whoops, I mean the affected consumers make a lot less money.
The third image in the “houseboat” series of posts that I’ve run on SOME social media outlets.
Beginning May 2, Bredemarket is on fewer social channels. The information in this post is not new.
If you were reading some of Bredemarket’s social channels, you may have seen a similar announcement on May 15, but that announcement listed the services where I was no longer posting.
If you were reading some of Bredemarket’s OTHER social channels, you may NOT have seen that announcement. Because those other social channels were the precise ones where I was no longer posting.
The May 15 social media account is a follow-up to a prior announcement from May 2, which said in part that I was “pausing activity on some Bredemarket social channels (and some related personal channels) that have no subscribers, exhibit no interest, or yield no responses.”
The second image in the “houseboat” series of posts that I’ve run on SOME social media outlets.
But rather than dwelling on the negative aspects of the social channels where Bredemarket has suspended activity, I’m going to concentrate on the social channels where Bredemarket is still active.
Of all the technologies I don’t know about, jobseeker technology is the most important. Between July 2000 and today, I’ve spent over 30 months searching for full-time employment. So it helps to know how employers search for potential employees.
And a lot of those 30-plus months have been filled with self-styled experts advising people “how to beat the applicant tracking system (ATS)” (as if there were only one) and “how to access the hidden job market” (because of course employers don’t want anyone to know that they’re searching for talent).
So when Melanie Woods of CGL Recruiting offered 20 jobseekers the chance to see how their LinkedIn Profile appears to users of LinkedIn Recruiter, I really wanted to win one of those 20 slots.
If I could boil all 7 tips down to 1, I’d emphasize that recruiters have limited time, and something a recruiter can understand in 0 seconds is much better than something that would take a recruiter 5 seconds to understand.
Instead of spending money on advertising, my team is going to be taking a different approach and it involves YOU #jobseekers!
Sounds like a plan, since a word-of-mouth testimonial converts more effectively than copy written by a marketing hack. (But what if the testimonial is FROM a marketing hack?)
The offer
She then described the offer.
We are going to be offering 20 free 20 minute sessions to job seekers who are currently out of work. During your session I will pull you up live in LinkedIn Recruiter and test your profile to show you where you are coming up in searches and where you are falling out. I will help you adjust your profile so you can appear in more searches for the jobs you are targeting.
What is LinkedIn Recruiter?
As you can probably guess, LinkedIn Recruiter is the premium-priced service that recruiters use to search LinkedIn for job candidates. The top-tier package (“contact Sales” for the price) includes:
Unlimited LinkedIn network access: Find and engage anyone on LinkedIn
150 InMail messages/month per license and bulk messaging
40+ advanced search filters, including “Open to work” and “More likely to respond”
Multi-user collaboration tools
Prepaid slots to rotate job postings in and out as needed
If your company is recruiting more than 4 positions a year, the high-end version of LinkedIn Recruiter could be the tool for you.
But when recruiters use LinkedIn Recruiter, they don’t look at a candidate’s LinkedIn profile—they look at the LinkedIn Recruiter view of the candidate’s profile, optimized for their purposes.
Hence Melanie was offering job applicants the opportunity to see how their profile appears to a recruiter. Valuable information to have.
What CGL Recruiting wanted in return
But remember that the lucky winners had to provide “advertising” to CGL Recruiting in return.
Here is where the advertising piece comes in….if you feel that the 20 minutes was helpful to you and your job search, we would appreciate you doing a review of our services on LinkedIn and one other social media platform (your choice), sharing how the time spent was useful to you and your job search.
Melanie and I met via Zoom early Monday afternoon Pacific Time (late afternoon Central Time), and I received a firehose of information during the 20-minute session. I’m not going to cover ALL the information she provided; instead, I’ll confine myself to the top 7 tips.
The first job on the LinkedIn profile is the most important.
Use all 5 “job title” slots.
Some employers AREN’T 1st and 2nd degree connections.
You can have 6 on-site job locations, not just 5.
Consider listing at least one college-related date.
Ampersands are bad.
Temperamental writers shouldn’t fall in love with pet phrases.
Tip 1: The first job on the LinkedIn profile is the most important
For my job search for a Senior Product Marketing Manager role, my Incode position is (maybe) more important than my current Bredemarket position. On my resume, I take care of this by listing Incode BEFORE Bredemarket. But because LinkedIn profiles are chronological, and Bredemarket is my current “employer,” I can’t reorder like that.
The LinkedIn Recruiter view of the profile doesn’t show all the positions, but only the top 3. And the first position takes great prominence.
LinkedIn Recruiter view of my LinkedIn profile, Monday, May 13, 2024, 1:30 pm PDT.
In the default view, the recruiter can’t see my fourth position (Strategic/Product Marketing Manager from 2015 to 2017), but only the first three. And only one of those three positions is product marketing-related.
Melanie zeroed in on my “Sole Proprietor” position, which tells a recruiter nothing about what I actually DO at Bredemarket. Sure the recruiter could click through and read about the marketing and writing services that Bredemarket provides…but recruiters have limited time.
I thought about her advice after the call, and for LinkedIn (and resume) purposes I’m changing my Bredemarket job title to “Product Marketing Consultant.” My work for my clients is all product/service-related, so the job title makes sense.
Now recruiters will see that two of my three most recent positions were product marketing-related, which makes me more attractive to the one position that I’m targeting.
One position? Wait a minute…
Tip 2: Use all 5 “job title” slots
That’s right. In my “job preferences,” my only listed job title was “Senior Product Marketing Manager.”
I can list up to 5.
Why not use all 5?
So now my job titles include the following:
Senior Product Marketing Manager
Product Marketing Manager
Marketing Content Manager
Global Product Marketing Manager
Product Marketing Consultant
So I have the position title from Incode, the consulting title from Bredemarket, two product marketing title variants, and a content marketing title for good measure (Bredemarket readers know why).
Now some people question why I’d list all these similar titles, since anyone who takes a few seconds can figure out that I’d be interested in a global product marketing manager position or whatever.
That’s the problem. Recruiters DON’T HAVE a few seconds. When hundreds or thousands of people apply for positions, recruiters need to get through the profiles as quickly as possible.
So Melanie wanted me to make her job easier.
Tip 3: Some employers AREN’T 1st and 2nd degree connections
You can control the visibility of your email address and your phone number on LinkedIn. While I don’t list a phone number on my LinkedIn profile, I do make my jobseeking email address (which is separate from my Bredemarket email address) visible. In fact, I configured my email address visibility for viewing by my 1st degree and 2nd degree connections.
But there was a fallacy in that tactic.
It became obvious in the session because Melanie (not a connection since her LinkedIn connections are maxed out) could NOT see my email address. Therefore, if recruiter Melanie wanted to contact me, she could ONLY contact me via InMail.
If you want ANY potential recruiter to see your email, increase its visibility to all connections. Obviously there are risks to this, so you need to judge what visibility is right for you. (Especially for phone numbers.)
Tip 4: You can have 6 on-site job locations, not just 5
While all of Bredemarket’s work is remote, and my previous work at Incode was remote, I’m not averse to on-site work. As long as it’s within driving distance.
To help local companies, I listed a selected five cities (the maximum) where I am available for on-site work:
Ontario, California, United States
San Bernardino, California, United States
Covina, California, United States
Pasadena, California, United States
Anaheim, California, United States
But Melanie pointed out that I didn’t need to list Ontario, since my profile already states that I live in Ontario. That freed up one slot to add another city. I chose to list Riverside, although I could have listed Brea or Fullerton or Industry or Pomona or Corona or many other cities. (LinkedIn, your on-site locations feature needs work.)
Tip 5: Consider listing at least one college-related date
Now let’s get into age discrimination talk.
If a company desires to discriminate against job applicants due to age, one effective way to do so is to look at the dates the applicants attended college. It’s pretty easy to quietly filter out the geezer applicants with no one the wiser.
For this reason I didn’t bother to add my college attendance dates to my LinkedIn profile. Why give the discriminating (in a negative way, not a positive way) firm the ammo they need to get the young, cheap workers they really want? (Of course those workers are inexperienced, but that’s another topic entirely.)
But Melanie pointed out one truth about companies that want to discriminate: if they don’t discriminate against you when they read your LinkedIn profile or resume, they can easily discriminate against you when they SEE you.
Oh, and there’s one more thing: if recruiters search for candidates based upon their graduation dates, profiles without graduation dates will never been seen by recruiters.
So I mulled over her advice.
I decided not to list the date that I started attending Cal State Fullerton’s MBA program.
And I decided not to list the date I graduated from Reed College.
I certainly didn’t list the date I started attending Reed College. (But I will confess that this song blasted from the Old Dorm Block. And I’ll also confess that I could lose the last 4 1/2 minutes.)
But I did pencil in my 1991 graduation date from Cal State Fullerton’s MBA program. Since my LinkedIn profile includes ALL my biometric positions going back to 1994, this isn’t a shocking revelation.
Tip 6: Ampersands are bad
LinkedIn profiles can include skills. I’ve listed near the limit of 50 skills, some of which are tied to particular positions, LinkedIn Learning courses, and other education and certifications.
Among many others, two of my listed skills are “identity & access management” and “sales & marketing management.”
Technology experts immediately see where this is going. So why didn’t I?
Melanie immediately noted that the ampersand character in those two skill descriptions can wreak havoc with some computerized systems.
I swapped out those skills for some new ones: identity and access management, and sales and marketing management, removing the problem.
Tip 7: Temperamental writers shouldn’t fall in love with pet phrases
Melanie’s biggest concern about my LinkedIn profile involved the very first sentence.
The one that appears below my profile, name, and preferred pronouns, but above my city of residence.
John Bredehoft LinkedIn profile, public view, Monday, May 13, 2024, 1:30 pm PDT.
Senior Product Marketing Manager in identity/technology who is expert in describing why customers benefit.
She read that sentence word for word.
“Senior Product Marketing Manager”? She liked that.
“In identity/technology”? I guess she liked that; at least she didn’t comment on it.
The rest of the sentence? Not so much.
Regular Bredemarket readers are familiar with the last three words of that sentence, and realize that every one of those three words is critically important. Why rather than what, customers rather than producers, and benefits rather than features. I’ve devoted a post (plus another post) to those three words. If I only had a few seconds to explain the importance of those three words…
Um, yet again, recruiters aren’t interested in taking a few minutes to read everything I have written about Simon Sinek’s Golden Circle, customer focus, or benefits. They don’t even want to watch a short reel on the topic.
So I have to rewrite the last part of that first sentence. As I write this post, I’m still mulling over alternatives.
CGL Recruiting’s expertise was highly beneficial
The 7 items above were just some of the tips that Melanie Woods imparted to me, all in the space of 20 minutes.
Again, the common theme is that recruiters have limited time, LinkedIn Recruiter lets them maximize that limited time, and jobseeker data also has to let recruiters maximize that limited time. If you can do something in 0 seconds, don’t take 5 seconds to do it.
I haven’t even explored some of the other features that CGL Recruiting offers, including Melanie Woods’ YouTube channel.
But the 20 minutes I spent with her were certainly valuable.
Now I just have to figure out another way to say “why customers benefit.”
I refrained from discussing this for a couple of days, but I was recently a victim of attempted financial identity fraud.
Well, SORT OF attempted identity fraud. I don’t know if this really counts, since I don’t know if the fraudster had my identity.
But the issue was resolved in less than 48 hours.
By the way, I have purposely changed the names of two of the companies I mention, to protect my PII. Which is a shame, because “Wildebeest Bank” went above and beyond in correcting the issue.
That doesn’t look right
Among its other services, Wildebeest Bank (not its real name) sends me an email whenever a purchase is made on my card, but my card is not present.
This is a fairly common occurrence. Among other things, my website, my business insurance, my business address, and my accounting software are all billed to my card.
But less than 48 hours ago, at 3:30 pm on Wednesday afternoon, I received an unexpected notice.
Your card was not present during a recent purchase
Your card was used to make a purchase at enron*publications us
We noticed your check card ending in 1234 was used to make a $8.48 purchase at enron*publications us today. The card wasn’t present at the time the purchase was made.
If you did not make this purchase, please call the nuber listed on the back of your card.
Log in to your account to review this transaction.
I didn’t recall making any $8.48 purchase, and once I looked up enron*publications us (not its real name), I realized that I definitely DIDN’T purchase anything from that company.
Before calling the bank, I double checked my account and found NO transaction for $8.48, even in a “pending” state.
So I called Wildebeest Bank
I called the number on the back of my card and connected with a woman in a call center who investigated why I got an email for a transaction that didn’t appear.
This is obviously not the Wildebeest Bank call center woman who helped me. But I’m sure she had a computer. By Earl Andrew at English Wikipedia – Own work, Public Domain, https://commons.wikimedia.org/w/index.php?curid=17793658
After accessing several internal systems, the woman discovered that the purchase was attempted, but declined. The fraudster had my card account number, but didn’t have the correct expiration date.
Frankly, I’m not even sure if the fraudster had my name. Did the fraudster just punch in 16 digits and hope they would work?
Anyway, after this conversation, the woman from Wildebeest Bank transferred me to the fraud department.
The Fraud Department
So my call was transferred to the Fraud Department.
The man at the Fraud Department advised me to cancel the card and get a new one.
I was wondering how long this would take, since one of bills was going to be charged to my card in the next two weeks, and I didn’t want any hiccup from a denied card purchase.
Anti-Fraud Man explained that if I could go to a Wildebeest Bank branch by the next day (Thursday), I could get a new card immediately.
“Could I go today?” I asked.
“Sure,” he replied.
It was about 3:50 pm by that time, or 20 minutes since I received the initial email.
So I drove to the bank
I hopped in my car, drove to a local bank branch, and went to a desk.
You may recall that I started Bredemarket in the fall of 2020, right in the middle of COVID. When I opened my account, the bank WOULDN’T let me go to my local bank branch and I had to open the account remotely. Since then I’ve been in the bank branch several times; it’s a nice place.
Anyway, the fraud department had already cancelled my compromised card, so the man at the bank branch only had to issue me a temporary card and guide me through its activation. This temporary card would last me until the new card arrived in the mail. It had the same card number as the new card so I could temporarily use it for purchases, but the permanent card would have a different expiration date and security code.
I could have provided the temporary card’s number, expiration date, and security code to the company that was going to bill me in two weeks, but I preferred to wait until I received the permanent card. I asked the man at the bank branch how long that would take.
“I can expedite it,” he said.
I get a present at Box 259
Less than 48 hours later, on Friday morning, I was notified that I had a package at my business address.
Bredemarket’s mailing address is 1030 N Mountain Ave #259, Ontario CA 91762-2114.
As I guessed, it was the permanent card, which I immediately activated and provided to the companies that auto-bill me via my card.
Here’s the short version:
My bank (“Wildebeest Bank”) notified me of a questionable “card not present” purchase (from “enron*publications us”) at 3:30 pm on Wednesday.
By 3:50 pm (20 minutes later), the bank told me that the attempted purchase was declined, but cancelled the bank card anyway.
By 4:15 pm (45 minutes later), I had a new temporary bank card.
By Friday at noon (less than 48 hours later), I had my permanent bank card.
So everyone be sure to bank at Wildebeest Bank. No confusion when you bank with them!
Vendors and researchers are paying a lot of attention to estimating ages by using a person’s face, and all of us are awaiting NIST’s results on its age estimation tests.
But before you declare dorsal hand features as the solution to age estimation, consider:
As the title states, the study only looked at females. No idea if my masculine hand features are predictive. (Anecdotally, more males work at tasks such as bricklaying that affect the hands, including the knuckle texture that was highlighted in the study.)
As the title states, the study only looked at people from India. No idea if my American/German/English/etc. hand features are predictive. (To be fair, the subjects had a variety of skin tones.)
The study only had 1454 subjects. Better than a study that used less than 20 people, but still not enough. More research is needed.
And even with all of that, the mean absolute error in age estimation was over 4 years.
It discussed both large language models and large multimodal models. In this case “multimodal” is used in a way that I normally DON’T use it, namely to refer to the different modes in which humans interact (text, images, sounds, videos). Of course, I gravitated to a discussion in which an image of a person’s face was one of the modes.
In this post I will look at LMMs…and I will also look at LMMs. There’s a difference. And a ton of power when LMMs and LMMs work together for the common good.
When Google announced its Gemini series of AI models, it made a big deal about how they were “natively multimodal.” Instead of having different modules tacked on to give the appearance of multimodality, they were apparently trained from the start to be able to handle text, images, audio, video, and more.
Other AI models are starting to function in a TRULY multimodal way, rather than using separate models to handle the different modes.
So now that we know that LLMs are large multimodal models, we need to…
…um, wait a minute…
Introducing the Large Medical Model (LMM)
It turns out that the health people have a DIFFERENT definition of the acronym LMM. Rather than using it to refer to a large multimodal model, they refer to a large MEDICAL model.
Our first of a kind Large Medical Model or LMM for short is a type of machine learning model that is specifically designed for healthcare and medical purposes. It is trained on a large dataset of medical records, claims, and other healthcare information including ICD, CPT, RxNorm, Claim Approvals/Denials, price and cost information, etc.
I don’t think I’m stepping out on a limb if I state that medical records cannot be classified as “natural” language. So the GenHealth.AI model is trained specifically on those attributes found in medical records, and not on people hemming and hawing and asking what a Pekingese dog looks like.
But there is still more work to do.
What about the LMM that is also an LMM?
Unless I’m missing something, the Large Medical Model described above is designed to work with only one mode of data, textual data.
But what if the Large Medical Model were also a Large Multimodal Model?
Rather than converting a medical professional’s voice notes to text, the LMM-LMM would work directly with the voice data. This could lead to increased accuracy: compare the tone of voice of an offhand comment “This doesn’t look good” with the tone of voice of a shocked comment “This doesn’t look good.” They appear the same when reduced to text format, but the original voice data conveys significant differences.
Rather than just using the textual codes associated with an X-ray, the LMM-LMM would read the X-ray itself. If the image model has adequate training, it will again pick up subtleties in the X-ray data that are not present when the data is reduced to a single medical code.
In short, the LMM-LMM (large medical model-large multimodal model) would accept ALL the medical outputs: text, voice, image, video, biometric readings, and everything else. And the LMM-LMM would deal with all of it natively, increasing the speed and accuracy of healthcare by removing the need to convert everything to textual codes.
A tall order, but imagine how healthcare would be revolutionized if you didn’t have to convert everything into text format to get things done. And if you could use the actual image, video, audio, or other data rather than someone’s textual summation of it.
Obviously you’d need a ton of training data to develop an LMM-LMM that could perform all these tasks. And you’d have to obtain the training data in a way that conforms to privacy requirements: in this case protected health information (PHI) requirements such as HIPAA requirements.
But if someone successfully pulls this off, the benefits are enormous.
You’ve come a long way, baby.
Robert Young (“Marcus Welby”) and Jane Wyatt (“Margaret Anderson” on a different show). By ABC TelevisionUploaded by We hope at en.wikipedia – eBay itemphoto informationTransferred from en.wikipedia by SreeBot, Public Domain, https://commons.wikimedia.org/w/index.php?curid=16472486.
I have worked with demand generation and growth marketing leader Gene Volfe, both as employees of the same company and as independent consultants working for other companies.
I asked him to review a short technology piece that I wrote, and he immediately found two gaps:
Bredemarket 
