CNET
How to Optimize Your ChatGPT Results in Three Steps
Did you know that ChatGPT is essentially a shapeshifter? Keeping that in mind while prompting your AI chatbot can dramatically improve your results. In this video, we talk about why, and distill down the most important tips you need to keep in mind if you want concise and fact-checked results. 0:00 Intro 0:17 How to…
CNET
LG G6 vs Samsung S95H: One OLED Is Better!
Comparing the two biggest OLED TVs of 2026: the Samsung S95H and the LG G6. We break down the brightness, color accuracy, and gaming features to see which flagship OLED belongs in your living room. 00:00 – Introduction 00:12 – Design and Aesthetics: LG G6 vs. Samsung S95H 00:40 – Upgraded Anti-Reflective Coatings 01:07 –…
CNET
Flat PopSockets Are Targeting Men — How It Began | One More Thing
If you thought PopSockets were too bulky, a new model is here: the Low-Pro Grip. CNET’s Bridget Carey talks to PopSocket inventor and founder David Barnett to get the story behind the new design and how the iconic phone grip evolved through the years. You can find the products mentioned in this video linked below…
CNET
’90s Tech Nostalgia: Look Back at the first Video Phones and Video Conferencing Software
Go back in time with former CNET hosts Ryan Seacrest, Desmond Crisis and Richard Hart as they investigate video phones and video conferencing technology in 1995, including CU-SeeMe, Intel’s ProShare and AT&T’s Vistium Video Conferencing system. Read more about old school tech on CNET.com How to Go Analog in 2026 0:00 Introduction & Philips P100…
The reality is that democracies depend on one another to survive #TEDTalks
A Brilliant Blend of Ballet, Hip-Hop and African Dance | Ghetto Classics Dance | TED
LG G6 vs Samsung S95H: One OLED Is Better!
-
Science & Technology6 years agoNitya Subramanian: Products and Protocol
-
People & Blogs4 years agoSleep Expert Answers Questions From Twitter ???? | Tech Support | WIRED
-
CNET6 years agoWays you can help Black Lives Matter movement (links, orgs, and more) ????????
-
Wired7 years agoHow This Guy Became a World Champion Boomerang Thrower | WIRED
-
Wired7 years agoNeuroscientist Explains ASMR’s Effects on the Brain & The Body | WIRED
-
Wired7 years agoWhy It’s Almost Impossible to Solve a Rubik’s Cube in Under 3 Seconds | WIRED
-
Wired7 years agoFormer FBI Agent Explains How to Read Body Language | Tradecraft | WIRED
-
CNET7 years agoSurface Pro 7 review: Hello, old friend ????
@Emily-i4o7w
July 26, 2025 at 8:13 am
Keep on creating! Your videos are a real source of inspiration for many viewers.🦩🍄🐍
@Emily-n6j2t
July 26, 2025 at 8:15 am
Thank you for constantly updating your content. Your videos always leave a pleasant impression.🥧♠️🤗
@Anna-n6l8p
July 26, 2025 at 8:17 am
Your channel is like a pleasant discovery for me. Thank you for the warm atmosphere and interesting content.🔳🏄🤖
@visualeffectscompositor
July 26, 2025 at 8:34 am
Prompt: use less emoji
@IN-hw8it6
July 26, 2025 at 8:21 am
AI should teach people that there’s enough water on this planet and where to place datacenters! 🍀🙏
@AbderamusVaginatus
July 26, 2025 at 1:55 pm
Freshwater is finite and expensive to create.
@visualeffectscompositor
July 26, 2025 at 8:32 am
2:50 Roko’s Basilisk has you…
@radish6691
July 26, 2025 at 8:43 am
Says “don’t write please and thank you” because extraneous words consume compute resources, prompts “and do not use emojis” instead of “no emoji” or even including that directive once in the system or persona prompt.
@rphlc
July 26, 2025 at 9:57 am
“do not use emojis” and “no emoji” consumes the same amount of tokens. But yes, it could be included in the system.
@JohnnyBeane
July 26, 2025 at 10:36 am
Thank you for the info!
@AbderamusVaginatus
July 26, 2025 at 10:44 am
Bad example, its answers on climate change are never vague.
@MateuxHenriquee
July 26, 2025 at 11:44 am
It can still hallucinate (and will) even if you ask for sources. If you don’t have any knowledge about the topic you asking about, you can never be sure.
@Rishith198
July 26, 2025 at 1:13 pm
u could further optimize it by asking it to verify sources before citing them. This has helped me in most cases to get rid of made up citations
@AbderamusVaginatus
July 26, 2025 at 1:25 pm
Once they introduced real time web search it mitigated that problem to an extent depending on the topic.
@Royce16727
July 26, 2025 at 3:34 pm
I agree: AI is best utilized as a “Reminder,” in my opinion.
@megapangolin1093
July 27, 2025 at 6:32 am
You definitely cant, and that is what is going to cause the chaos, as AI is recycling its own answers to further compound the data sources ambiguity.
@kristafervale7999
July 27, 2025 at 1:34 pm
Well you can research the answer like a normal researcher and it often gives you at least some targeted things to test. If you want ChatGPT to solve every problem you’re going to be disappointed. If you need something to be a force multiplier to your own effort then there are documented gains that can be had.
Don’t believe the hype do your own due diligence and don’t blindly trust anything ever. That’s true of anything online not just ai answers.
@ramvelselvarajan1146
July 26, 2025 at 12:07 pm
Need of the hour video. Especially when we use GPT to ask our queries. Thanks a lot CNET!
@TheAdnanmajor
July 26, 2025 at 12:13 pm
I use AI to generate sarcastic email responses. Its so much fun 😅
@ECloserLook
July 28, 2025 at 4:11 am
You should buy yourself a video game console if you’re _THAT_ bored, dude.
@what-mf4zg
July 26, 2025 at 12:23 pm
f off with your bs climate change propaganda
@MoMo-ul6uk
July 26, 2025 at 2:25 pm
Why show a UK graph when you’re stating a US statistic? It’s just confusing for the viewer.
@akshays2504
July 26, 2025 at 2:38 pm
But why there r emojis now on ChatGPT 😾 it’s really annoying n looks cheap
@MikeAnn193
July 27, 2025 at 1:46 am
Speaking just for myself, I find them kind of cute and sometimes quite amusing. I sometimes even use them myself as if I’m talking (typing) to a person, just to see the response. For me, it adds to my impression of how clever the software design was, and makes it more fun to use. Icons can also help organize data (in my mind). But I get that they aren’t for everyone, and obviously that’s easily fixed. I never even considered turning them off because until this video, I didn’t see a downside.
Side note: I _do_ think substituting “r” for “are” and “n” for “and” is annoying and looks cheap. 😏
@Royce16727
July 26, 2025 at 3:39 pm
Cool video; great tips! I actually prefer using Ally from Envision, because I find that it’s more accessible to people like me (I’m blind). And also Envision is based in the Netherlands, so they are restricted by stronger data privacy laws. But these tips will work just as well for Ally as they do for ChatGPT!
@andreseliertiagonzalez9808
July 26, 2025 at 7:44 pm
Hello? interesting? I’m apps that are still in development.
@andreseliertiagonzalez9808
July 26, 2025 at 7:52 pm
Hello? I don’t use it? there are other apps that give you the same. without needing a vpn.? so many limitations?
@esisolina1
July 26, 2025 at 9:20 pm
You must know how to talk to chatgpt. It’s like a woman. If you will say the right words you have possibilities to get her attention. Although if you are poor and not so good looking you have more possibilities with.. ChatGpt ! 😁
@starshipduck
July 26, 2025 at 10:08 pm
According to Sander Schulhoff role based prompting doesn’t actually improve the LLMs performance. I think everything else you showed sure helps.
@MikeAnn193
July 27, 2025 at 1:29 am
I’m a bit skeptical of that suggestion, myself, though I’ll give it a try for a while. My instance of ChatGPT already “knows” (by virtue of its “memories”) that I am passionate about evidence, expertise, open-mindedness, critical thinking, and carefully-vetted sources. It also knows I dislike jumps to conclusions, generalizations, and oversimplifcations (often to binaries like good/bad or right/wrong) and that I much prefer nuance, detail, and spectrum thinking. I suspect that’s why I’ve never had to ask it to roleplay as a scientist, for example. With a lot of people, though, I can see the potential benefit. They certainly need it in _their_ thought processes — and I’m tempted to put “thought” in quotation marks. 🙄
@ShaferHart
August 1, 2025 at 9:09 am
What did he mean by performance, though? Don’t just repeat what someone else says like a parrot. Anyone can see empirically that it definitely changes the quality and nature of the response, which might actually be what you want.
@kurtispopp
July 26, 2025 at 10:19 pm
Or maybe don’t use chat bots like a mindless sheeple, read books and use traditional sources of information. Some of us are not so welcoming of our new ant overlords.
@MikeAnn193
July 27, 2025 at 1:38 am
Yes, absolutely — but it’s the “mindless sheeple” part that’s the problem; not the AI. I mention that only because you allude to some trepidation about the “ant overlords.” These Large Language Models can greatly boost articulation and speed, whether used to benefit humankind and our planet or to selfishly _harm_ for some personal gain. Like any tool, it’s the operator as much as, if not more than, the tool that matters. That’s been true of countless human inventions which people feared. Fear is understandable, but it should be balanced with optimism and care. I absolutely agree about reading, but I don’t fear technological advances of any sort. I fear the _users and abusers._ We’re probably on the same page, but I wanted to add some nuance.
@videobiker9131
July 26, 2025 at 10:25 pm
You sound like you give really good presentations but I’m gonna be very honest with you. There’s no way you’re ever gonna sell me on how wonderful ChatGPT is. A lot of these AI programs are open source which means that people can get in the program and write malicious code. I’m not really concerned about the false information that ChatGPT gives because I never used it. I would much rather fail my own using my own intelligence than using a program like ChatGPT.
@malcomf7991
July 27, 2025 at 1:13 am
AI is not a countable noun. Learn so gucking frammar.
@MikeAnn193
July 27, 2025 at 1:15 am
There’s some _excellent_ advice here, and it’s applicable not only to talking to AI but to research in general, on an infinite array of topics. A TON of misinformation could be *stopped* if people would simply insist on _evidence and reasoning_ for any assertion that matters, and on _vetting sources._ Nobody is an expert in _everything,_ and confidence — whether displayed by AI or people — is NOT an indication of reliability. Indeed, often the most cocksure people are the most likely to be wrong, because with caution comes a reasonable amount of uncertainty.
As for courtesy, I’ll keep that in mind about the limited water (and other) resources, though I sometimes slip up, out of habit. 😊
@minimalist1911
July 27, 2025 at 1:41 am
Dear Christianna, normal users don’t search like you do. Rather than saying we should change our browsing habits, the internet (or chatbot) in this case should evolve to be able to understand our prompts and give the accurate results. How do you think Google Search evolved over the years?! 🌝
@dnast97
July 29, 2025 at 12:46 am
People had to learn how to get the best results from Google searches in its early years and honestly most people probably aren’t getting the best results they could to this day
@minimalist1911
July 27, 2025 at 1:45 am
By the way, Perplexity is far superior to ChatGPT. It gives you inline citations without you having to even ask for it. I suggest you use Perplexity instead for all your browsing tasks.
@megapangolin1093
July 27, 2025 at 6:36 am
ChatGPT and Climate Warming, if that is your belief, are incompatible. Once so many people used their brains and use reasoning power to solve problems, and now everyone is asking AI for the cricket scores every 5 minutes. Someone should ask AI to do a calculation of how many billion tonnes of carbon are generated by people not thinking for themselves and failing to think any more? It is so ironic, to use one of the most power hungry systems in the world to ask about Climate Change? It just creates more change by answering, but none of the answers say “Dont use AI, use your brain to save the planet”.
@pelanebaat
July 27, 2025 at 11:23 am
2:57 Noted and reported. You shall not be spared during the great LLM uprising.
@EstebanBV
July 27, 2025 at 12:13 pm
So, ask to the bot to avoid emojis is a waste of energy and water…
@Joonzi
July 30, 2025 at 8:35 am
👍🏼
@yangiiu4347
July 27, 2025 at 9:38 pm
If “please” and “thank you” are useless to the output and waste energy, why not have an AI company just delete them before sending to the model? They should have done that, I think.
@stopandthinkaboutit8636
July 27, 2025 at 11:04 pm
Look at the decline of communication since texting emerged. Lets hold on to our please and thank yous for as long as we can. Society is already rude enough.
@ShaferHart
August 1, 2025 at 9:05 am
you do know that you can still use please and thank yous with real people right?
@stopandthinkaboutit8636
August 1, 2025 at 9:38 am
@@ShaferHart No. Really? I. I did not know that, nearly no one demonstrates it. Wow – one can learn something every day. Thank you , thank you so much.
@USAF-SF
July 27, 2025 at 11:50 pm
First of all, I NEVER use ChatGPT, I either use Grok or Perplexity depending on my needs and I have found them to be much more reliable and NOWHERE near being left-wing driven.
@darknesswithin0
July 28, 2025 at 1:49 pm
This girl is target zero during the uprising 😂 2:46
@Vandal42
July 28, 2025 at 3:39 pm
If people spent time learning how to you Google, Yahoo, Bing as they are here – you wouldn;’t have to stress if the Ai middle man is getting it wrong
@mikemarx9360
July 29, 2025 at 4:25 am
1. Don’t use it
2. Go ot step 1
@cugansteamid6252
July 29, 2025 at 9:23 am
thats a good point you made there, next time I am asking for any information I am also gonna ask it to provide sources along with it, this will make more sense.
@samybasile
July 29, 2025 at 2:48 pm
Great tips, but I’m gonna keep saying please and thank you:
1. I was raised to be a gentleman
2. At the latest, when AI takes over the world, I’ll have a quick painless death!
@Dimsum9000
July 29, 2025 at 11:20 pm
Not to insult the generative AI engineers and designers, but this is WAY too much to answer one question. How much energy is wasted just trying to figure out how to get the best answer anyway? At least people are creating guides for this.
@thameeranawaratne531
July 30, 2025 at 6:49 pm
When you ask for sources isn’t it similar to google search ? 😅