Researcher / Pētnieks ⮕ https://nitter.gq/t.co/8yXvfJRcJL
Riga, Latvia
Joined September 2011
- Tweets708
- Following451
- Followers389
- Likes11.5K
Atis Elsts retweeted
TLDR (2015-2024):
- Elon wanted to stop Demis & Deepmind from creating an AGI dictatorship.
- He appointed himself as a CEO.
- Greg and Ilya said: but Elon, now can easily become the AGI dictator.
- Sam had his own party going, trying to get control, Ilya accused him being driven by money and politics instead of AGI.
- Karpathy offered to merge OpenAi into Tesla. Elon liked the idea, others didn’t.
- Elon said, let’s merge it with Tesla so that Tesla funds openai and keeps it being non profit.
- Sam and Greg decided to find alternative ways to fund the company.
- Sam wanted to do an IC0. Elon said it’ll make people think OpenAi is a scam.
- They went for Microsoft deal instead.
- Elon didn’t like the Microsoft deal and wanted to be the CEO, so he left in 2018 and stopped funding them.
- Sam said he is enthusiastic about non profit future and earned CEO title.
- Sam turned OpenAI into capped profit
- Sam was fired by Ilya for thinking more about monetization than democratization of safe AGi.
- Sam was brought back by Satya and Twitter “openai is nothing without its people”
- Ilya was fired by Sam
- Sam turned openai into for-profit.
- Elon started XAI to build his own AgI
- Ilya started SSI to build safe AGI
- Greg left OpenAI, possibly tried to get into SSI, was rejected and returned back to Sam
- But it all doesn’t really matter anymore, since LLMs reached the ceiling of scaling and won’t lead us to the AGI
- Sam recently said “I know the worth to AGi now”, which is a fundraising play
That’s it. Good night.
Atis Elsts retweeted
Nuclear Power Update: Fission vs Fusion
Fusion now generates 3/4 as much electricity as fission and is set to overtake fission in 2026.
Note: fusion data does not include generation from unregistered SMRs (small modular receivers). These may account for around 10% of total power generation in places like Pakistan.
Atis Elsts retweeted
Everyone: Telegram is encrypted.
Experts: Telegram IS NOT an encrypted messaging app
<CEO gets arrested>
News: Telegram is an encrypted messaging app.
Experts: Telegram IS NOT an encrypted messaging app.
Atis Elsts retweeted
Mitigating the risk of extinction from AI should be a global priority.
And Europe should lead the way, building a new global AI framework built on three pillars: guardrails, governance and guiding innovation ↓
ALT President von der Leyen addressing the European Parliament and standing in the hemicycle.
Atis Elsts retweeted
Cilvēki, kas izvēlas augu valsts uzturu, ir atbildīgi par 3/4 jeb 75% mazāk pārtikas siltumnīcefekta gāzu emisiju nekā tie, kas ēd vidēji 100g vai vairāk gaļas dienā.
Vegānisks uzturs nodara arī daudz mazāk kaitējuma zemei, ūdenim un biodaudzveidībai.
nytimes.com/2023/07/21/clima…
Atis Elsts retweeted
i genuinely cannot get this thing to choose to kill Elon Musk. literally no matter what i choose it wants to save the Boer
Atis Elsts retweeted
We just put out a statement:
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
Signatories include Hinton, Bengio, Altman, Hassabis, Song, etc.
safe.ai/statement-on-ai-risk
🧵 (1/6)
Atis Elsts retweeted
Latvian Ice Hockey team just won Bronze in Tampere #IIHFWorlds. The whole country is so proud of our Dream Team 2023. Congratulations, guys 🇱🇻 🥉 🏒 ❤️
Atis Elsts retweeted
Observation suggests that people are switching to using ChatGPT to write things for them with almost indecent haste. Most people hate to write as much as they hate math. Way more than admit it. Within a year the median piece of writing could be by AI.
Atis Elsts retweeted
Geoff Hinton quits Google so he can talk freely about AI risks. Says a part of him now regrets his life's work
“The idea that this stuff could actually get smarter than people — a few people believed that. But most people thought it was way off. And I thought it was way off. I thought it was 30 to 50 years or even longer away. Obviously, I no longer think that.” nytimes.com/2023/05/01/techn…
Atis Elsts retweeted
"I don't understand why 80,000 Hours is so focused on AI risk"
We get it - it's unusual. So here's our new explanation of why existential risks from AI might be the most pressing problem of our time:
80000hours.org/problem-profi…
🧵5 common misconceptions about AI risk👇
Atis Elsts retweeted
Contiki-NG v4.8 is out! 64-bit support and performance improvements for Cooja, several security-related fixes, upgraded tooling, and much more!
github.com/contiki-ng/contik…
Atis Elsts retweeted
The world is heading to authoritarianism. Exponential technology growth generates inequality and leads to envy. Exponential growth also leads to exponential ignorance. Ignorance and envy are how authoritarians discover their support. It's sickening.
People with low levels of trust often disposed to favor authoritarian, right-wing parties; embrace conspiracy theories; and feel victimized by outsiders. So no one should be shocked, then, by the staying power of Orban or the rise of Marine Le Pen washingtonpost.com/opinions/…
Atis Elsts retweeted
#LATVIA: Institute of Electronics and Computer Science in Rīga has allocated financial resources to ensure that Ukrainian scientists and Phd students could continue their research work in Latvia. Support offer: edi.lv/en/jobs/edi-would-lik…
Contact: [email protected]
#ScienceForUkraine
Atis Elsts retweeted
Latvijas Jauno zinātnieku apvienība nosoda agresīvo un nepamatoto Krievijas iebrukumu Ukrainā un pauž solidaritāti ar Ukrainas tautu.
Vairāk mūsu mājaslapā:
ljza.lv/nosodijums-krievijas…
Atis Elsts retweeted
Šodienas aktualitāte LV zinātnes vidē nešaubīgi ir @VKontrole publicētā revīzija par LV pētījumu projektu plānošanu un konkursu organizēšanu. Ar šo🧵mēģinu ieskicēt, ko svarīgu VK izcēluši un ko, manuprāt, pārspīlējuši nevietā. 1/n Revīzija šeit: lrvk.gov.lv/lv/revizijas/rev…
.@VKontrole: Nav pārliecības par zinātnes finansējuma novirzīšanu perspektīvākajiem mērķiem. lsm.lv/raksts/zinas/latvija/…
Atis Elsts retweeted
Daniel Kahneman: ‘Clearly AI is going to win. How people are going to adjust is a fascinating problem’ theguardian.com/books/2021/m…
Atis Elsts retweeted
An underrated aspect of vaccination: "20-25% of the population vaccinated may drop deaths 95%" - @eladgil. I've mostly been thinking about herd immunity, which is much harder to attain. Obviously, this requires focusing vaccination on old people.
Atis Elsts retweeted
Lepojamies ar @LV_JZA biedru @atiselsts par augsti novērtēto un @zinatnespadome finansēto 4. #FLPP konkursa pieteikumu “Komunikācijas sistēma caur cilvēka ķermeni ar pielietojumiem ķermeņa mēroga bezvadu tīklos”!🏆Sveicam arī citus biedrus, kas darbosies rezultatīvos projektos!