This is the problem.
3 Likes
The āparagraph from a sentence fragmentā is my most used feature. I am either brief or excessively wordy, no in between. Working when efficiency matters, I chose brief. If an email could be answered with yes or no, I would do so, politely, with a little words as necessary.
But after enough complaints that I was being short or dismissive because I was not treating every email as my next magnum opus. I relented and started using AI to spin my āyesā into a soliloquy that would make Shakespeare weep to look back upon his own works.
it makes me sick but workplace āpositivityā has grown to the point where if Iām not stroking everybody gently with the written word, they cry that theyāre offended.
5 Likes
Iāve been getting a ton of mileage from Gemini for my current vacation.
I had an itinerary of my tripās timed tickets and such for seeing sites but itās poorly formatted. Scribbled notes and times. I popped it into Gemini and cleaned it up
I have a bunch of recommendations for my trip in texts and email. I grabbed all the raw text as context and had Gemini curate the list by neighborhoods and categories (site, restaurant, etc.).
Then I mixed in my itinerary and had it cluster the recommendations around them so I had lists of things to do around my plans.
Finally, I had it recommend transit between activities, with recommendations for walking if the area is nice with things to see vs. subways vs. cabs for convenience.
Havenāt fact checked everything but itās mostly been processing provided content. Everything itās churned out has been helpful so far.
5 Likes
Yes I use AI heavily for random personal junk.was looking for cybersecurity conferences and used gpt to make a list I can review
I was actually thinking that someone needs to create an AI tour-guide app. These are the types of apps that I think we will see popping up soon. Until AI can reason I think its best use will be for answering specific questions; about specific things.
2 Likes
You have to double check it then too unfortunately.
Iāve been using Geminiās canvas mode instead of writing documents by hand recently and itās magical. Even if I have it generate a single big document I then go through it line by line suggesting point fixes. Itās way faster than writing.
I also do a more line by line generation, or use it to guide me through filling in templated documents.
1 Like
The more AI that gets shoved into my work day the more I realize the spend isnāt worth the output. Itās a bet lose in most aspects. I donāt get how this bubble hasnāt popped yet..
1 Like
Bubble hasnāt popped because itās only gotten started. The software phase is still in the middle innings and the hardware game hasnāt even really begun.
At the rate things are moving we will have regular rolling brown outs across the east coast in a few years due to AI over utilizing the grid. Decanters are going up so rapidly especially in northern VA (the hub of the internet where 80% of all internet traffic intersects) the companies that are buying the hardware are at the point investors are now asking when do we make the money back? The bubble is real and with the economy stuttering and stagnation continuing even with the federal cuts. The bubble is bound to degrade at minimum .
I know Iām always doom and gloom about this stuff but at some point reality has to meet expectations and it just isnāt. Tesla is already finally meeting reality and flailing as expected itās only a matter of time till companies realize they canāt keep buy hardware and that will topple all the dominoes.
4 Likes
That is why it is a bubble. How long this phase will last is the big question. No one knows the result of the current infrastructure build-out . While the current AI is truly amazing, it is still incapable of reasoning. Without reasoning, the āhallucinations (aka: incorrect answers) will never be resolved.
Will the current AI be eclipsed by larger, more complex LLMās employing tens of thousands of NVIDIA processors; lead to more intelligence? Will the entry of custom chips like those from Broadcom result in new, innovative models and capabilities that allow for new reasoning and super-intelligence? No one knows. That is why Meta and others are paying individual researchers outlandish salaries to come work for them and not the competition. The first to the prize will be the big winner. so the incentive to invest is great.
My view is that sooner or later, there will be a paper published in some obscure AI journal that essentially says that the current LLMās are not providing better responses as the number of processors increase. Probably it will report the opposite; that more processors actually yield worse results. That is what would be a catalyst for the AI hardware bubble bursting. That is the point where software will emerge as the new frontier of AI.
1 Like
The hardware Iām talking about isnāt processors for ai data centers. Itās the humanoid robots, self driving everything, and autonomous manufacturing thatās coming.
Ahh yes, the vaunted self-driving cars and humanoid robots. You are a Telsa investor then? 
I am holding my breath in anticipation.
1 Like
Iām not an investor. Just someone who is willing to see whatās obvious. Youāre gonna be holding your breath for a long time because like I said this is just starting. Letās talk again in 5-10 years.
Speed of progress in both of these areas will be determined by the speed of innovation in AI. As noted above; that is the big question. Are we on the brink of super-intelligence? Or, will it be more of an iterative process that will evolve over time?
Self driving may never get past regulations which is the real hurdle. Reality is while self driving trucking would probably be safer killing the driving gig economy and trucking overnight would cripple the US economy almost worse than nvidia failing due to a hardware bubble crash. I see regulators stepping in to stop innovation to protect their base in the end and that stalling self driving and AI robotics for decades which will kill the leading companies in those spaces today. Also Tesla stepping backwards recently on self driving was a death throw that may not be survivable for that company with how itās had so much negativity recently as well.
Not politics just noting recent events and speculating on how people respond to threats to their well beings. I see an anti clanker (robots) movement coming as strong as the anti immigrant one has been in recent yearsā¦
1 Like
I do see manufacturing changing but Iām seeing more localized manufacturing becoming the norm. I could see AI playing a part there but in the end I see making things shifting over the next decade to being hyper localized
1 Like
Regardless AI is currently very dependent on those Datacenters and I just canāt square that circle yetā¦
1 Like
All your points sound eerily similar to what they were saying about the internet when I was in high school in the mid 90s. Itāll never get through regulations, uses too much electricity, just a novelty, everyone will revolt against it, y2k will destroy us, wonāt change anything, etc etc etc.
The folks who were entrenched in a horse world said the same thing about cars when they came out too.
On the hardware end once quantum hardware becomes relavent itās gonna get real nutz.
For every car or internet thereās a split strike conversion strategy, the disks reading device or flying cars that promised massive change but never could or where just a Ponzi scheme. Crypto claims to be one of those amazing things but currently the only ones able to utilize crypto effectively as a currency are either scammers, drug trade or other criminal groups.
AI might change the world but could also prove in the short term to be way to expensive and not meet the short term outlook of investors resulting in eventually fizzling out like many other great ideas that never came to be because either market conditions didnāt accept it, people decidedly feared it or it just wasnāt technically possible at the time.
We could have lived in a world where Nicola Tesla developed wireless power generations before us but fear and greed never allowed it to happen. We could have had air ships all over the place but sadly one massive incident struck fear and killed the entire industry leaving only 4 operational in all of the US todayā¦
Iām skeptical and will remain as such till o see actual value. All I see today is a way to pump stock tickers and folks getting emotionally attached to very expensive chat bots and allot of justification for more allot of slop that needs allot of people to shape it into anything actually usableā¦
3 Likes