Rachel Vandernick (@VandernickR) header
Rachel Vandernick (@VandernickR) profile picture

Digital marketing for internet obsessed brands. Freelancer for hire. Occasional adjunct & writer. ✡️ Tweeting marketing & freelancing.

Cape Coral, FL
Joined September 2012
Attn freelancers with invoice anxiety-- I worked with an influencer today who charges $25k for **ONE** TikTok video. So breathe and hit send on those invoices with confidence because it's ~literally~ fine.
18
244
52
1,433
Rachel Vandernick retweeted
Lily nails it here! The fundamental issue isn't just that LLMs hallucinate, it's that AI companies have completely failed to educate users about what these tools actually are & how they work. What users need to know: LLMs are pattern completion engines, not knowledge databases. They generate outputs based on statistical patterns in their training data. They don't "know" facts or "understand" context, they predict what words should come next based on patterns they've seen before. This is why they hallucinate with such confidence. The model generates answers the same way whether the information is accurate or completely fabricated. They cannot tell the difference! Remember, these are language models, not fact-checking models. This fundamental limitation makes AI dangerous for medical advice, legal guidance, financial decisions, or any high-stakes situation where accuracy matters. But AI companies are NOT teaching or alerting users to this, they're marketing these tools as "almost AGI" & racing toward growth while leaving a trail of misinformation, lies and misplaced trust.. As Lily points out, this isn't the user's fault. There's no exam required to use ChatGPT. The responsibility (and accountability) lies with AI companies who prioritize competitive pressure > user safety. What we need from AI companies & policymakers: - Mandatory user education on what LLMs can and cannot do - Clear warning labels on high-risk topics (medical, legal, financial advice) - Actual regulation (& not the theater "regulate us!" we see from Altman) What YOU can do right now: ✅ Always verify AI outputs: especially for medical, legal, financial, or factual claims. (Use multiple sources) 🚫 Never use AI as your only source for high-stakes decisions 📢 Push back on AI hype: including journalists who promote these tools without discussing limitations or risks 🎓 Teach others: share what you learn about how these tools actually work 🚀 Demand better: from AI companies prioritizing growth over user safety This tech has fundamental limitations that can't be fixed with additional training. Until companies are honest about that, we need each other more than ever. ❤️ Real expertise matters. Human judgment matters. Community knowledge-sharing matters. Don't let AI replace the relationships & networks that actually keep us informed and safe. 🦸‍♀️ The antidote to AI hype isn't rejecting the technology; it's building stronger communities of informed users who can call out BS, share what actually works, and protect each other from harm. (More on how we can do this together coming soon 👀 - Follow for updates) Stay skeptical! Stay connected! +Keep having these conversations
I keep seeing posts all over the place where people are sharing how they use AI (LLMs) to do this or that - but in reality, LLMs can’t actually do the thing they are suggesting. In my opinion, the main problem here - which has been a problem since day 1 of these tools becoming available to the public - is that they rarely say when they don’t know something or can’t do something, but they usually give a confident answer anyway. It can feel exciting for someone new or unfamiliar with AI to ask ChatGPT to “act like a heart surgeon” or “pretend to be a tax expert” and get a confident, plausible response. This is especially true given the media frenzy around AI - it seems fair for most people to assume that this must be the revolutionary, life-changing technology we were promised. But in reality, those answers might have been entirely hallucinated and full of inaccurate or even dangerous information. The user often doesn’t know the difference. And I’m not saying I’m immune to this, I’ve found myself alarmed by inaccurate answers in ChatGPT many times only after using Google to fact-check its response. I don’t blame the general public for being misled by LLMs; it’s not like there is some exam required to use this technology that we are now learning can be genuinely dangerous. The real fault lies with AI companies for failing to teach users what these tools can and can’t do (though that would obviously clash with their extreme growth goals). But now that the cat is out of the bag - it’s very worrying to think how much incorrect information is circulating around the internet and in society, and how many people are making mistakes ranging from taking a harmless wrong turn to buying the wrong medicine for their needs. They’ll never do this, but it would be great if these AI companies took bigger steps to: - train users on the best way to use these tools (or not) - flag when certain topics are dangerous - improve the tools’ ability to say when they aren’t confident in their response (or have some type of confidence meter??) We could also, IDK, have laws regulating this stuff. ¯\_(ツ)_/¯ I also highly recommend following @BritneyMuller to learn more about the best ways to use LLMs, plus the risks and dangers of using it for the wrong things.
12
61
7
295
28,289
Rachel Vandernick retweeted
The Eagles emergency third QB today
25
444
24
5,127
355,670
The mental gymnastics, reverse causality, and revisionist history that far leftists engage in to normalize jihad is no accident, and I’m no longer pretending that it is
1,034
3,796
719
14,331
1,209,117
Rachel Vandernick retweeted
If you’re Jewish and not with your family or congregation today, remember that you are part of a community that is not bound to a physical location. An eternally unified people, impossible to divide, committed to a set of undying principles. Shana Tova.
164
548
40
6,831
281,539
This is my favorite ad of the year and it’s not even close.
If you talk sh!t, you better back it up. Watch what happens when real Cuervo sh!t talkers put their taste to the test and go one round in the Octagon with a UFC Interim Champ, Tom Aspinall.

120
Rachel Vandernick retweeted
Today, the FTC finalized a new rule banning fake reviews. One interesting part: it expressly prohibits buying and selling fake followers. But it goes beyond followers to include "any metrics used by the public to make assessments of ... social media influence."
38
289
78
1,282
140,551
Rachel Vandernick retweeted
General Lebron James crossing the Delaware
288
16,198
665
190,475
5,675,151
Rachel Vandernick retweeted
Ethical issues aside for the moment, this is a trap for unwary advertisers. The FTC is clear that all endorsements must reflect the honest beliefs of the endorser, and that explicitly includes virtual/AI influencers. If you use a deepfake/AI creator and script them to say how much they love a product, without a clear and conspicuous disclosure that the person is not real and has not used the product, then you are breaking the law—they cannot possibly have used the product, so their endorsement cannot reflect their honest belief.
This Post is from an account that no longer exists. Learn more
10
73
11
299
82,496
A week and a half later @Upwork I’ve gotten no response from anyone and several more emails about fraud positions from applicants. Someone is scamming applicants on your site under my brand and your team isn’t doing anything to address it.
Replying to @VandernickR
Hi Rachel, sorry to hear about this! It appears you've been able to open a support ticket with our support team. You can see any updates here: spr.ly/6016d0eAI
1
1
166
.@Upwork someone is impersonating me and my company through false job postings on your website. The automated chat support was useless. Please help!
1
1
78
Rachel Vandernick retweeted
What's the risk of using copyrighted music in organic brand content? Last week, Sony sued Marriott hotels for copyright infringement for including unlicensed music in IG and TikTok posts. If music is part of your brand's (or its influencers') social strategy, read this 👇
10
54
12
225
77,373
Okay hear me out— Minus equity, what is being described is freelancing ✨
This quoted post is unavailable.
65
Rachel Vandernick retweeted
requesting PTO at work is so embarrassing. "hi boss permission to enjoy my life for 3 days?"
347
23,033
2,339
234,611
11,623,479
My unpopular take on the AI Overview situation: if the SEO industry were as obsessed with the actual user behavior of seeking as it is the delivery method of content, we would be in a collectively smarter, more agile place. (And other hot takes below): prdaily.com/what-googles-ai-…
1
100
A message from Bill Maher to the anti-Israel activists. Once again, Bill is spot on! 🇮🇱❤️
171
2,054
196
8,784
522,185
Rachel Vandernick retweeted
Again, brands, please do not post celeb pics on your IGs just because they wore your stuff! Yesterday, shoe brand Stuart Weitzman was sued for copyright infringement for posting a pic of Kendall Jenner on its Instagram.
13
55
10
349
180,125
Rachel Vandernick retweeted
Check out this post sponsored by Fenty Skin. The creator used IG's "Paid partnership" tool and included both #ad and #sponsored. But the NAD recently determined that this post did not comply with the FTC's endorsement disclosure requirements. Three takeaways:
13
30
16
155
94,650
ten years as an antiquarian, hearing nothing but "my relative kept these untouched in a box their whole life but they died, and I don't want this stuff" has radicalized me. annotate the books. use the china. wear the clothes until they wear out. your life should not be a museum
This Post is from an account that no longer exists. Learn more
131
11,299
321
63,863
2,356,697
Rachel Vandernick retweeted
We can break down silos to boost results on social 📈 #RaganDisney
1
4
229