DavidGarcia

@DavidGarcia@feddit.nl

• ⁨Comment⁩ on ⁨AI Search Engines Are Confidently Wrong Too Often.⁩ ⁨⁨1⁩ ⁨week⁩ ago⁩:

  no search has been garbage for more than 10 years now, years before any usable AI model even existed. AI articles just made it worse.

  SEO and relentless profit optimization of search engines without regard for UX is what is to blame in my book.

• ⁨Comment⁩ on ⁨AI Search Engines Are Confidently Wrong Too Often.⁩ ⁨⁨2⁩ ⁨weeks⁩ ago⁩:

  what I like about using AI for search is that even if it’s wrong, it’s at least pointing me in the right direction.

  whereas if I use a modern search engine I find nothing.

  I remember when search engines would actually give you what you were looking for. those were the days. I would take that over AI search any time.

  When you could actually get useful exact matches and semantic matches too.

  nowadays search engines are more like:

  “if page contains one of the stem words in the query, then return the page as top result; ignore all other words in the query or what order they are in”

  someone at google probably got 1000000 billion dollars for this innovative “optimization”

• ⁨Comment⁩ on ⁨AI Search Engines Are Confidently Wrong Too Often.⁩ ⁨⁨2⁩ ⁨weeks⁩ ago⁩:

  I would be interested to see a comparison between AI search and classic search. “How often does it find what I am looking for?” and “How often is that info confidently wrong?” compared.

  Classic search is so garbage these days, I can never find what I am looking for.

• ⁨Comment⁩ on ⁨Steam Deck fans, you're in for a good one - AMD's FSR 3 frame generation is now available on Valve's handheld thanks to a new Decky plugin⁩ ⁨⁨1⁩ ⁨month⁩ ago⁩:

  the tragedy of frame gen. works the worst when you need it the most.

• ⁨Comment⁩ on ⁨Intel Fails To Achieve Gaudi AI "$500 Million" Revenue Target, Showing Slacking Performance In The AI Segment.⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  I think the problem is not many companies want to invest in AI accelerators since it’s not clear if AI will even make them any money.

  If you just buy GPUs, it’s much more likely that you can repurpose them if AI turns out to be a bust.

• ⁨Comment⁩ on ⁨Wikipedia is under assault: rogue users keep posting AI generated nonsense⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  AI is the nuke of the World Information War 1

• ⁨Comment⁩ on ⁨Reddit communities will require permission while going private or switching to NSFW⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  why anyone is still on reddit is beyond me. it’s so garbage at this point that it makes me angry any time I so much as see it referenced

• ⁨Comment⁩ on ⁨LG TVs start showing ads on screensavers⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  I notice a major trend of all the largest companies majorly shooting themselves in the stock with turbo greed, even more greed than normal. I feel like we’re on the verge of a stock market slaughter.

  “maximum ads”, “X basic necessity as an investment”, “price gouge everything via cartels/monopolies” and “X as a service” are the death nails of the current economic order.

• ⁨Comment⁩ on ⁨LG TVs start showing ads on screensavers⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  watch them download ads over ultrasound mesh networks

• ⁨Comment⁩ on ⁨Due to AI fakes, the “deep doubt” era is here⁩ ⁨⁨6⁩ ⁨months⁩ ago⁩:

  reality is deeply unserious

• ⁨Comment⁩ on ⁨Black Myth: Wukong shows very clearly Valve are selling a lot of Steam Decks⁩ ⁨⁨6⁩ ⁨months⁩ ago⁩:

  Imma deck you in the schnoz 😃

• ⁨Comment⁩ on ⁨Solar modules deployed in France in 1992 still provide 79.5% of original output power⁩ ⁨⁨9⁩ ⁨months⁩ ago⁩:

  Those are way more interesting than this single installation with -0.66%:

  “Another more recent study carried out by the US Department of Energy’s National Renewable Energy Laboratory (NREL) on 1,700 American sites totaling 7.2 GW of power, showed a median degradation of around -0.75%/year. Moveover, another research focused on 4,300 residential installations in operation in Europe and used different data processing methodologies. Depending on the methods, a median loss of -0.36% to -0.67%/year was obtained.”

  Also the article doesn’t mention standard deviation or IQR of the drop, or what efficiency they originally had. Which is the minimum of information I would expect.

• ⁨Comment⁩ on ⁨Meta’s AI is summarizing some bizarre Facebook comment sections⁩ ⁨⁨9⁩ ⁨months⁩ ago⁩:

  I always thought it would be cool to have an AI that only shows you maximally diverse comments on social media. Like if you have 90% of all people saying the same thing and 10% saying something different, you get 2 comments summarizing each opinion.

  Especially on Reddit everything is buried under the dribble of the hivemind. It would be cool if you could only extract the actually interesting viewpoints.

• ⁨Comment⁩ on ⁨Modder Boosts Steam Deck to a Bountiful 32GB RAM⁩ ⁨⁨1⁩ ⁨year⁩ ago⁩:

  I’m having homebrew psp flashbacks