Tag: run

  • Will future computers run on human brain cells? Breaking ground on new field of \’organoid intelligence\’

    جانز ہاپکنز یونیورسٹی کے محققین کے مطابق، انسانی دماغ کے خلیوں سے چلنے والا ایک \”بائیو کمپیوٹر\” ہماری زندگی کے اندر تیار کیا جا سکتا ہے، جو اس طرح کی ٹیکنالوجی سے جدید کمپیوٹنگ کی صلاحیتوں کو تیزی سے وسعت دینے اور مطالعے کے نئے شعبے تخلیق کرنے کی توقع رکھتے ہیں۔

    ٹیم آج جرنل میں \”آرگنائڈ انٹیلی جنس\” کے لیے اپنے منصوبے کا خاکہ پیش کرتی ہے۔ سائنس میں فرنٹیئرز.

    جانز ہاپکنز بلومبرگ سکول آف پبلک ہیلتھ اینڈ وائٹنگ سکول آف انجینئرنگ میں ماحولیاتی صحت سائنسز کے پروفیسر تھامس ہارٹنگ نے کہا، \”کمپیوٹنگ اور مصنوعی ذہانت ٹیکنالوجی کے انقلاب کو آگے بڑھا رہی ہے لیکن وہ ایک حد تک پہنچ رہی ہیں۔\” \”بائیو کمپیوٹنگ کمپیوٹیشنل پاور کو کمپیکٹ کرنے اور ہماری موجودہ تکنیکی حدود کو آگے بڑھانے کے لیے اس کی کارکردگی کو بڑھانے کی ایک بہت بڑی کوشش ہے۔\”

    تقریباً دو دہائیوں سے سائنس دانوں نے انسانی یا جانوروں کی جانچ کے بغیر گردوں، پھیپھڑوں اور دیگر اعضاء پر تجربہ کرنے کے لیے چھوٹے آرگنائڈز، لیبارٹری میں تیار کردہ بافتوں کا استعمال کیا ہے جو مکمل طور پر بڑھے ہوئے اعضاء سے مشابہت رکھتے ہیں۔ ابھی حال ہی میں ہارٹنگ اور جانس ہاپکنز کے ساتھی دماغی آرگنائڈز کے ساتھ کام کر رہے ہیں، نیوران اور دیگر خصوصیات کے ساتھ قلم کے نقطے کے سائز کے مدار میں کام کر رہے ہیں جو سیکھنے اور یاد رکھنے جیسے بنیادی افعال کو برقرار رکھنے کا وعدہ کرتے ہیں۔

    ہارٹنگ نے کہا کہ \”یہ تحقیق کھولتا ہے کہ انسانی دماغ کیسے کام کرتا ہے۔\” \”کیونکہ آپ نظام میں ہیرا پھیری شروع کر سکتے ہیں، ایسے کام کر سکتے ہیں جو آپ اخلاقی طور پر انسانی دماغ سے نہیں کر سکتے۔\”

    ہارٹنگ نے 2012 میں دماغی خلیات کو فنکشنل آرگنائڈز میں اکٹھا کرنا شروع کیا اور انسانی جلد کے نمونوں کے خلیات کو برانن سٹیم سیل جیسی حالت میں دوبارہ پروگرام کیا گیا۔ ہر آرگنائیڈ میں تقریباً 50,000 خلیے ہوتے ہیں، جو کہ پھل کی مکھی کے اعصابی نظام کے سائز کے ہوتے ہیں۔ اب وہ ایسے دماغی آرگنائڈز کے ساتھ مستقبل کا کمپیوٹر بنانے کا تصور کرتا ہے۔

    ہارٹنگ نے کہا کہ اس \”حیاتیاتی ہارڈویئر\” پر چلنے والے کمپیوٹرز اگلی دہائی میں سپر کمپیوٹنگ کی توانائی کی کھپت کے مطالبات کو کم کرنا شروع کر سکتے ہیں جو تیزی سے غیر پائیدار ہوتے جا رہے ہیں۔ اگرچہ کمپیوٹرز انسانوں کے مقابلے میں اعداد اور ڈیٹا پر مشتمل حسابات پر تیزی سے عمل کرتے ہیں، دماغ پیچیدہ منطقی فیصلے کرنے میں زیادہ ہوشیار ہوتے ہیں، جیسے کتے کو بلی سے بتانا۔

    ہارٹنگ نے کہا کہ \”دماغ اب بھی جدید کمپیوٹرز سے بے مثال ہے۔ فرنٹیئر، کینٹکی میں جدید ترین سپر کمپیوٹر، 600 ملین ڈالر، 6,800 مربع فٹ کی تنصیب ہے۔ صرف پچھلے سال جون میں، یہ پہلی بار کسی ایک انسانی دماغ کی کمپیوٹیشنل صلاحیت سے تجاوز کر گیا — لیکن اس سے دس لاکھ گنا زیادہ استعمال ہو رہا ہے۔ توانائی.\”

    ہارٹنگ نے کہا کہ آرگنائڈ انٹیلی جنس سسٹم کو ماؤس کی طرح سمارٹ بنانے میں کئی دہائیاں لگ سکتی ہیں۔ لیکن دماغی آرگنائڈز کی پیداوار کو بڑھا کر اور انہیں مصنوعی ذہانت سے تربیت دے کر، وہ ایک ایسے مستقبل کی پیشین گوئی کرتا ہے جہاں بائیو کمپیوٹرز اعلیٰ کمپیوٹنگ کی رفتار، پروسیسنگ پاور، ڈیٹا کی کارکردگی، اور ذخیرہ کرنے کی صلاحیتوں کی حمایت کرتے ہیں۔

    ہارٹنگ نے کہا، \”کسی بھی قسم کے کمپیوٹر کے مقابلے میں کسی چیز کا ہدف حاصل کرنے میں ہمیں کئی دہائیاں لگیں گی۔\” \”لیکن اگر ہم اس کے لیے فنڈنگ ​​پروگرام بنانا شروع نہیں کرتے تو یہ بہت زیادہ مشکل ہو جائے گا۔\”

    جانز ہاپکنز کی ماحولیاتی صحت اور انجینئرنگ کی اسسٹنٹ پروفیسر لینا سمرنووا نے کہا کہ آرگنائیڈ انٹیلی جنس نیورو ڈیولپمنٹل عوارض اور نیوروڈیجنریشن کے لیے منشیات کی جانچ کی تحقیق میں بھی انقلاب لا سکتی ہے۔

    سمرنوفا نے کہا، \”ہم عام طور پر تیار شدہ عطیہ دہندگان کے دماغی آرگنائڈز کا موازنہ کرنا چاہتے ہیں بمقابلہ آٹزم کے ساتھ عطیہ دہندگان کے دماغ کے آرگنائڈز\”۔ \”جو ٹولز ہم بائیولوجیکل کمپیوٹنگ کے لیے تیار کر رہے ہیں وہی ٹولز ہیں جو ہمیں جانوروں کو استعمال کیے بغیر یا مریضوں تک رسائی کیے بغیر، آٹزم کے لیے مخصوص نیورونل نیٹ ورکس میں ہونے والی تبدیلیوں کو سمجھنے کی اجازت دیں گے، اس لیے ہم ان بنیادی میکانزم کو سمجھ سکتے ہیں کہ مریضوں کو یہ ادراک کیوں ہوتا ہے۔ مسائل اور خرابیاں۔\”

    آرگنائڈ انٹیلی جنس کے ساتھ کام کرنے کے اخلاقی مضمرات کا اندازہ لگانے کے لیے، سائنسدانوں، حیاتیاتی ماہرین اور عوام کے اراکین کا ایک متنوع کنسورشیم ٹیم کے اندر شامل کیا گیا ہے۔

    جانز ہاپکنز کے مصنفین میں شامل ہیں: برائن ایس کیفو، ڈیوڈ ایچ گراسیا، کیو ہوانگ، اٹزی ای مورالس پینٹوجا، بوہاؤ تانگ، ڈونلڈ جے زیک، سنتھیا اے برلینیک، جے لوماکس بوائیڈ، ٹموتھی ڈی ہیرس، ایرک سی جانسن، جیفری کاہن، بارٹن ایل پالہمس، جیسی پلاٹکن، ​​الیگزینڈر ایس زالے، جوشوا ٹی ووگلسٹین، اور پال ایف ورلی۔

    دیگر مصنفین میں شامل ہیں: بریٹ جے کاگن، کورٹیکل لیبز کے؛ ایلیسن آر. موتری، یونیورسٹی آف کیلیفورنیا سان ڈیاگو کے؛ اور یونیورسٹی آف لکسمبرگ کے جینس سی شومبورن۔



    >Source link>

    >>Join our Facebook page From top right corner. <<

  • Will future computers run on human brain cells? Breaking ground on new field of \’organoid intelligence\’

    جانز ہاپکنز یونیورسٹی کے محققین کے مطابق، انسانی دماغ کے خلیوں سے چلنے والا ایک \”بائیو کمپیوٹر\” ہماری زندگی کے اندر تیار کیا جا سکتا ہے، جو اس طرح کی ٹیکنالوجی سے جدید کمپیوٹنگ کی صلاحیتوں کو تیزی سے وسعت دینے اور مطالعے کے نئے شعبے تخلیق کرنے کی توقع رکھتے ہیں۔

    ٹیم آج جرنل میں \”آرگنائڈ انٹیلی جنس\” کے لیے اپنے منصوبے کا خاکہ پیش کرتی ہے۔ سائنس میں فرنٹیئرز.

    جانز ہاپکنز بلومبرگ سکول آف پبلک ہیلتھ اینڈ وائٹنگ سکول آف انجینئرنگ میں ماحولیاتی صحت سائنسز کے پروفیسر تھامس ہارٹنگ نے کہا، \”کمپیوٹنگ اور مصنوعی ذہانت ٹیکنالوجی کے انقلاب کو آگے بڑھا رہی ہے لیکن وہ ایک حد تک پہنچ رہی ہیں۔\” \”بائیو کمپیوٹنگ کمپیوٹیشنل پاور کو کمپیکٹ کرنے اور ہماری موجودہ تکنیکی حدود کو آگے بڑھانے کے لیے اس کی کارکردگی کو بڑھانے کی ایک بہت بڑی کوشش ہے۔\”

    تقریباً دو دہائیوں سے سائنس دانوں نے انسانی یا جانوروں کی جانچ کے بغیر گردوں، پھیپھڑوں اور دیگر اعضاء پر تجربہ کرنے کے لیے چھوٹے آرگنائڈز، لیبارٹری میں تیار کردہ بافتوں کا استعمال کیا ہے جو مکمل طور پر بڑھے ہوئے اعضاء سے مشابہت رکھتے ہیں۔ ابھی حال ہی میں ہارٹنگ اور جانس ہاپکنز کے ساتھی دماغی آرگنائڈز کے ساتھ کام کر رہے ہیں، نیوران اور دیگر خصوصیات کے ساتھ قلم کے نقطے کے سائز کے مدار میں کام کر رہے ہیں جو سیکھنے اور یاد رکھنے جیسے بنیادی افعال کو برقرار رکھنے کا وعدہ کرتے ہیں۔

    ہارٹنگ نے کہا کہ \”یہ تحقیق کھولتا ہے کہ انسانی دماغ کیسے کام کرتا ہے۔\” \”کیونکہ آپ نظام میں ہیرا پھیری شروع کر سکتے ہیں، ایسے کام کر سکتے ہیں جو آپ اخلاقی طور پر انسانی دماغ سے نہیں کر سکتے۔\”

    ہارٹنگ نے 2012 میں دماغی خلیات کو فنکشنل آرگنائڈز میں اکٹھا کرنا شروع کیا اور انسانی جلد کے نمونوں کے خلیات کو برانن سٹیم سیل جیسی حالت میں دوبارہ پروگرام کیا گیا۔ ہر آرگنائیڈ میں تقریباً 50,000 خلیے ہوتے ہیں، جو کہ پھل کی مکھی کے اعصابی نظام کے سائز کے ہوتے ہیں۔ اب وہ ایسے دماغی آرگنائڈز کے ساتھ مستقبل کا کمپیوٹر بنانے کا تصور کرتا ہے۔

    ہارٹنگ نے کہا کہ اس \”حیاتیاتی ہارڈویئر\” پر چلنے والے کمپیوٹرز اگلی دہائی میں سپر کمپیوٹنگ کی توانائی کی کھپت کے مطالبات کو کم کرنا شروع کر سکتے ہیں جو تیزی سے غیر پائیدار ہوتے جا رہے ہیں۔ اگرچہ کمپیوٹرز انسانوں کے مقابلے میں اعداد اور ڈیٹا پر مشتمل حسابات پر تیزی سے عمل کرتے ہیں، دماغ پیچیدہ منطقی فیصلے کرنے میں زیادہ ہوشیار ہوتے ہیں، جیسے کتے کو بلی سے بتانا۔

    ہارٹنگ نے کہا کہ \”دماغ اب بھی جدید کمپیوٹرز سے بے مثال ہے۔ فرنٹیئر، کینٹکی میں جدید ترین سپر کمپیوٹر، 600 ملین ڈالر، 6,800 مربع فٹ کی تنصیب ہے۔ صرف پچھلے سال جون میں، یہ پہلی بار کسی ایک انسانی دماغ کی کمپیوٹیشنل صلاحیت سے تجاوز کر گیا — لیکن اس سے دس لاکھ گنا زیادہ استعمال ہو رہا ہے۔ توانائی.\”

    ہارٹنگ نے کہا کہ آرگنائڈ انٹیلی جنس سسٹم کو ماؤس کی طرح سمارٹ بنانے میں کئی دہائیاں لگ سکتی ہیں۔ لیکن دماغی آرگنائڈز کی پیداوار کو بڑھا کر اور انہیں مصنوعی ذہانت سے تربیت دے کر، وہ ایک ایسے مستقبل کی پیشین گوئی کرتا ہے جہاں بائیو کمپیوٹرز اعلیٰ کمپیوٹنگ کی رفتار، پروسیسنگ پاور، ڈیٹا کی کارکردگی، اور ذخیرہ کرنے کی صلاحیتوں کی حمایت کرتے ہیں۔

    ہارٹنگ نے کہا، \”کسی بھی قسم کے کمپیوٹر کے مقابلے میں کسی چیز کا ہدف حاصل کرنے میں ہمیں کئی دہائیاں لگیں گی۔\” \”لیکن اگر ہم اس کے لیے فنڈنگ ​​پروگرام بنانا شروع نہیں کرتے تو یہ بہت زیادہ مشکل ہو جائے گا۔\”

    جانز ہاپکنز کی ماحولیاتی صحت اور انجینئرنگ کی اسسٹنٹ پروفیسر لینا سمرنووا نے کہا کہ آرگنائیڈ انٹیلی جنس نیورو ڈیولپمنٹل عوارض اور نیوروڈیجنریشن کے لیے منشیات کی جانچ کی تحقیق میں بھی انقلاب لا سکتی ہے۔

    سمرنوفا نے کہا، \”ہم عام طور پر تیار شدہ عطیہ دہندگان کے دماغی آرگنائڈز کا موازنہ کرنا چاہتے ہیں بمقابلہ آٹزم کے ساتھ عطیہ دہندگان کے دماغ کے آرگنائڈز\”۔ \”جو ٹولز ہم بائیولوجیکل کمپیوٹنگ کے لیے تیار کر رہے ہیں وہی ٹولز ہیں جو ہمیں جانوروں کو استعمال کیے بغیر یا مریضوں تک رسائی کیے بغیر، آٹزم کے لیے مخصوص نیورونل نیٹ ورکس میں ہونے والی تبدیلیوں کو سمجھنے کی اجازت دیں گے، اس لیے ہم ان بنیادی میکانزم کو سمجھ سکتے ہیں کہ مریضوں کو یہ ادراک کیوں ہوتا ہے۔ مسائل اور خرابیاں۔\”

    آرگنائڈ انٹیلی جنس کے ساتھ کام کرنے کے اخلاقی مضمرات کا اندازہ لگانے کے لیے، سائنسدانوں، حیاتیاتی ماہرین اور عوام کے اراکین کا ایک متنوع کنسورشیم ٹیم کے اندر شامل کیا گیا ہے۔

    جانز ہاپکنز کے مصنفین میں شامل ہیں: برائن ایس کیفو، ڈیوڈ ایچ گراسیا، کیو ہوانگ، اٹزی ای مورالس پینٹوجا، بوہاؤ تانگ، ڈونلڈ جے زیک، سنتھیا اے برلینیک، جے لوماکس بوائیڈ، ٹموتھی ڈی ہیرس، ایرک سی جانسن، جیفری کاہن، بارٹن ایل پالہمس، جیسی پلاٹکن، ​​الیگزینڈر ایس زالے، جوشوا ٹی ووگلسٹین، اور پال ایف ورلی۔

    دیگر مصنفین میں شامل ہیں: بریٹ جے کاگن، کورٹیکل لیبز کے؛ ایلیسن آر. موتری، یونیورسٹی آف کیلیفورنیا سان ڈیاگو کے؛ اور یونیورسٹی آف لکسمبرگ کے جینس سی شومبورن۔



    >Source link>

    >>Join our Facebook page From top right corner. <<

  • Will future computers run on human brain cells? Breaking ground on new field of \’organoid intelligence\’

    جانز ہاپکنز یونیورسٹی کے محققین کے مطابق، انسانی دماغ کے خلیوں سے چلنے والا ایک \”بائیو کمپیوٹر\” ہماری زندگی کے اندر تیار کیا جا سکتا ہے، جو اس طرح کی ٹیکنالوجی سے جدید کمپیوٹنگ کی صلاحیتوں کو تیزی سے وسعت دینے اور مطالعے کے نئے شعبے تخلیق کرنے کی توقع رکھتے ہیں۔

    ٹیم آج جرنل میں \”آرگنائڈ انٹیلی جنس\” کے لیے اپنے منصوبے کا خاکہ پیش کرتی ہے۔ سائنس میں فرنٹیئرز.

    جانز ہاپکنز بلومبرگ سکول آف پبلک ہیلتھ اینڈ وائٹنگ سکول آف انجینئرنگ میں ماحولیاتی صحت سائنسز کے پروفیسر تھامس ہارٹنگ نے کہا، \”کمپیوٹنگ اور مصنوعی ذہانت ٹیکنالوجی کے انقلاب کو آگے بڑھا رہی ہے لیکن وہ ایک حد تک پہنچ رہی ہیں۔\” \”بائیو کمپیوٹنگ کمپیوٹیشنل پاور کو کمپیکٹ کرنے اور ہماری موجودہ تکنیکی حدود کو آگے بڑھانے کے لیے اس کی کارکردگی کو بڑھانے کی ایک بہت بڑی کوشش ہے۔\”

    تقریباً دو دہائیوں سے سائنس دانوں نے انسانی یا جانوروں کی جانچ کے بغیر گردوں، پھیپھڑوں اور دیگر اعضاء پر تجربہ کرنے کے لیے چھوٹے آرگنائڈز، لیبارٹری میں تیار کردہ بافتوں کا استعمال کیا ہے جو مکمل طور پر بڑھے ہوئے اعضاء سے مشابہت رکھتے ہیں۔ ابھی حال ہی میں ہارٹنگ اور جانس ہاپکنز کے ساتھی دماغی آرگنائڈز کے ساتھ کام کر رہے ہیں، نیوران اور دیگر خصوصیات کے ساتھ قلم کے نقطے کے سائز کے مدار میں کام کر رہے ہیں جو سیکھنے اور یاد رکھنے جیسے بنیادی افعال کو برقرار رکھنے کا وعدہ کرتے ہیں۔

    ہارٹنگ نے کہا کہ \”یہ تحقیق کھولتا ہے کہ انسانی دماغ کیسے کام کرتا ہے۔\” \”کیونکہ آپ نظام میں ہیرا پھیری شروع کر سکتے ہیں، ایسے کام کر سکتے ہیں جو آپ اخلاقی طور پر انسانی دماغ سے نہیں کر سکتے۔\”

    ہارٹنگ نے 2012 میں دماغی خلیات کو فنکشنل آرگنائڈز میں اکٹھا کرنا شروع کیا اور انسانی جلد کے نمونوں کے خلیات کو برانن سٹیم سیل جیسی حالت میں دوبارہ پروگرام کیا گیا۔ ہر آرگنائیڈ میں تقریباً 50,000 خلیے ہوتے ہیں، جو کہ پھل کی مکھی کے اعصابی نظام کے سائز کے ہوتے ہیں۔ اب وہ ایسے دماغی آرگنائڈز کے ساتھ مستقبل کا کمپیوٹر بنانے کا تصور کرتا ہے۔

    ہارٹنگ نے کہا کہ اس \”حیاتیاتی ہارڈویئر\” پر چلنے والے کمپیوٹرز اگلی دہائی میں سپر کمپیوٹنگ کی توانائی کی کھپت کے مطالبات کو کم کرنا شروع کر سکتے ہیں جو تیزی سے غیر پائیدار ہوتے جا رہے ہیں۔ اگرچہ کمپیوٹرز انسانوں کے مقابلے میں اعداد اور ڈیٹا پر مشتمل حسابات پر تیزی سے عمل کرتے ہیں، دماغ پیچیدہ منطقی فیصلے کرنے میں زیادہ ہوشیار ہوتے ہیں، جیسے کتے کو بلی سے بتانا۔

    ہارٹنگ نے کہا کہ \”دماغ اب بھی جدید کمپیوٹرز سے بے مثال ہے۔ فرنٹیئر، کینٹکی میں جدید ترین سپر کمپیوٹر، 600 ملین ڈالر، 6,800 مربع فٹ کی تنصیب ہے۔ صرف پچھلے سال جون میں، یہ پہلی بار کسی ایک انسانی دماغ کی کمپیوٹیشنل صلاحیت سے تجاوز کر گیا — لیکن اس سے دس لاکھ گنا زیادہ استعمال ہو رہا ہے۔ توانائی.\”

    ہارٹنگ نے کہا کہ آرگنائڈ انٹیلی جنس سسٹم کو ماؤس کی طرح سمارٹ بنانے میں کئی دہائیاں لگ سکتی ہیں۔ لیکن دماغی آرگنائڈز کی پیداوار کو بڑھا کر اور انہیں مصنوعی ذہانت سے تربیت دے کر، وہ ایک ایسے مستقبل کی پیشین گوئی کرتا ہے جہاں بائیو کمپیوٹرز اعلیٰ کمپیوٹنگ کی رفتار، پروسیسنگ پاور، ڈیٹا کی کارکردگی، اور ذخیرہ کرنے کی صلاحیتوں کی حمایت کرتے ہیں۔

    ہارٹنگ نے کہا، \”کسی بھی قسم کے کمپیوٹر کے مقابلے میں کسی چیز کا ہدف حاصل کرنے میں ہمیں کئی دہائیاں لگیں گی۔\” \”لیکن اگر ہم اس کے لیے فنڈنگ ​​پروگرام بنانا شروع نہیں کرتے تو یہ بہت زیادہ مشکل ہو جائے گا۔\”

    جانز ہاپکنز کی ماحولیاتی صحت اور انجینئرنگ کی اسسٹنٹ پروفیسر لینا سمرنووا نے کہا کہ آرگنائیڈ انٹیلی جنس نیورو ڈیولپمنٹل عوارض اور نیوروڈیجنریشن کے لیے منشیات کی جانچ کی تحقیق میں بھی انقلاب لا سکتی ہے۔

    سمرنوفا نے کہا، \”ہم عام طور پر تیار شدہ عطیہ دہندگان کے دماغی آرگنائڈز کا موازنہ کرنا چاہتے ہیں بمقابلہ آٹزم کے ساتھ عطیہ دہندگان کے دماغ کے آرگنائڈز\”۔ \”جو ٹولز ہم بائیولوجیکل کمپیوٹنگ کے لیے تیار کر رہے ہیں وہی ٹولز ہیں جو ہمیں جانوروں کو استعمال کیے بغیر یا مریضوں تک رسائی کیے بغیر، آٹزم کے لیے مخصوص نیورونل نیٹ ورکس میں ہونے والی تبدیلیوں کو سمجھنے کی اجازت دیں گے، اس لیے ہم ان بنیادی میکانزم کو سمجھ سکتے ہیں کہ مریضوں کو یہ ادراک کیوں ہوتا ہے۔ مسائل اور خرابیاں۔\”

    آرگنائڈ انٹیلی جنس کے ساتھ کام کرنے کے اخلاقی مضمرات کا اندازہ لگانے کے لیے، سائنسدانوں، حیاتیاتی ماہرین اور عوام کے اراکین کا ایک متنوع کنسورشیم ٹیم کے اندر شامل کیا گیا ہے۔

    جانز ہاپکنز کے مصنفین میں شامل ہیں: برائن ایس کیفو، ڈیوڈ ایچ گراسیا، کیو ہوانگ، اٹزی ای مورالس پینٹوجا، بوہاؤ تانگ، ڈونلڈ جے زیک، سنتھیا اے برلینیک، جے لوماکس بوائیڈ، ٹموتھی ڈی ہیرس، ایرک سی جانسن، جیفری کاہن، بارٹن ایل پالہمس، جیسی پلاٹکن، ​​الیگزینڈر ایس زالے، جوشوا ٹی ووگلسٹین، اور پال ایف ورلی۔

    دیگر مصنفین میں شامل ہیں: بریٹ جے کاگن، کورٹیکل لیبز کے؛ ایلیسن آر. موتری، یونیورسٹی آف کیلیفورنیا سان ڈیاگو کے؛ اور یونیورسٹی آف لکسمبرگ کے جینس سی شومبورن۔



    >Source link>

    >>Join our Facebook page From top right corner. <<

  • Nothing’s Phone (2) will run on Qualcomm’s Snapdragon 8 series

    Nothing’s presence stood in stark contrast to OnePlus at this year’s MWC. Whereas Carl Pei’s old company, OnePlus, put on a flashy launch event for a concept device yesterday, Nothing has largely used the event to meet with vendors and other big names in the industry.

    We sat down with Pei yesterday, fittingly in a meeting room inside the Qualcomm booth. Among other things, the CEO confirmed that the Nothing (2) will be powered by the Snapdragon 8 series chip — not that the topic was ever up for much debate.

    The conversation — which touches on the U.S. and India markets and the smartphone industry at large — began with a quick exchange about the OnePlus concept phone. Earlier teasers of the device drew comparisons to Nothing’s first handset, the Phone (1). “People were tagging me,” Pei says, while acknowledging that the device’s illuminated cooling liquid is a distinct approach from his phone’s Glyph lighting scheme.

    CP: There’s [a] company at MWC called UniHertz. A very small company. They made a carbon copy of the Phone (1). I’m gonna go check it out.

    TC: Congratulations. That’s a rite of passage.

    We recently worked on our vision statement. It’s, “We want to make tech fun again.” So, if we can inspire our industry to start experimenting more, that still helps with the vision.

    There’s experimenting and there’s copying. Are you going to go be litigious?

    No. They’re so small. That doesn’t really help anybody.

    The smartphone market was on the decline before the pandemic. It got a bump from 5G and then declined again. At some point 6G will happen, but I don’t see something like foldables driving the market in a profound way. Will the market rebound? And if so, how?

    The smartphone market grew initially because there was a really innovative product that was useful to customers. Now it’s starting to shrink, because my phone is good enough. Why should I upgrade?

    It’s also too expensive to upgrade every two years.

    Yeah. I think it’s natural that unless there’s some really useful innovation, it’s going to steadily decline. Having said that, it’s still a really big industry. It’s still very lucrative from a business perspective.

    So, you don’t see anything on the immediate horizon that will profoundly drive sales?

    Not profoundly. For us, we’re trying to do more and more as we build up our engineering capabilities. We weren’t able to do too much on the software side the first year. But now that we have our own team, we can start doing more and more, but it’s not going to be game changing. I don’t think in the next generation, but we’re steadily introducing something new and useful, I think. But I don’t see that iPhone moment on the horizon moment anymore. Not in the next two to three years.

    \"smartphone\"

    Image Credits: Brian Heater

    There may not be one. That was the ur-moment for phones.

    Yeah. I think something around AI might be the next iPhone, but I haven’t really figured out the application.

    You’re talking about AI on the phone?

    No, AI as a technology to help people in general.

    Obviously there have been some applications for AI on the phone, mostly for photography.

    Yeah. But AI on the smartphone has only given us like 10% improvement. The picture quality is 10% better, but it doesn’t change how we interact with technology.

    What are your feelings on foldables?

    I personally think foldables are supply chain-driven innovation and not consumer insights.

    In terms of having that initial breakthrough and then building devices around it?

    Think about it from a supply chain perspective. Somebody invents OLED, and they can make a lot of money, because it’s a great technology. Then after a few years, a lot more companies make that, so they need to lower their prices. So they need to figure out what else they can sell at a higher margin. They develop flexible OLEDs, which they can sell at a higher price. Then they go and pitch it to the smartphone vendors. “We have this nice foldable, please use it.”

    In Samsung’s case, they developed the technology in-house and built the device around it. As far as I can tell, they still own the vast majority of that market right now.

    Foldables [don’t] come from solving customer pain points, but it’s getting better and better. That’s great, but it’s probably not going to be something we’re going to look at.

    \"\"

    Image Credits: Brian Heater

    So, we shouldn’t expect a foldable from Nothing?

    No.

    There’s some talk of 6G at this year’s event, but it still feels pretty far off. What’s your sense on where the technology is?

    I don’t know anything. I think for us it’s not where we need to leave. We just follow what the industry trend is on the network side. We need to lead where we can be different. If we ask our users why they bought the earphones and smartphone, the number one reason is design. So, how can we get stronger on that? Not just the hardware design, but to also adapt that to our software? So everything feels holistic. And then, how can we make our technology more useful? We have the Glyph interface on the back that looks quite nice, quite interesting. But we’ve got a lot of feedback that it wasn’t that functional. How can we build more and more functionality that’s actually usable?

    In this case, you adopted a technology specifically for design reasons and then started looking for use cases?

    It’s a mix, because we had a roadmap for features of the Glyph interface. But that roadmap has been executed on very slowly, because we didn’t have the engineering needed.

    How big is the team now?

    We’re 410. Still very small.

    How many are in engineering?

    Probably 350.

    The vast majority.

    Yeah, unfortunately, the smartphone is a very complicated product, so you need a lot of engineers.

    Is Nothing profitable?

    As a hardware business in a competitive market we have now sold 1,000,000 Nothing products and counting in just over two years.  Whilst Nothing won’t be profitable in its early days, our revenue grew tenfold from 2021 to 2022.

    What can you tell me about the Phone (2)?

    We’re going to be using the [Snapdragon] 8 series. Earlier, I said it was going to be a premium device. But we’ve never officially acknowledged whether it’s Qualcomm or MediaTek.

    Was that ever up for debate? MediaTek is good, but Snapdragon is the clear choice for most flagships.

    Qualcomm has been a really good partner. From the very beginning, when we were starting the business, there was a chipset shortage, and they gave us allocation at a good price. Today it’s different. There’s now an abundance of chipsets in the market, but they’ve been a strong supplier. And their product’s not bad, so we never really considered another option, especially for a more premium device.

    Is there anything else you can share on the Nothing (2) front?

    We’ll have a much stronger focus on software. With our engineering in-house, there’s a lot more we can do. We have a more robust roadmap, both in terms of design and how we can make it more useful.

    Given the amount of time and money that went into R&D, will the design be similar to the Nothing (1)?

    In terms of cost it will probably be similar.

    And in terms of design language?

    That I can’t really comment on.

    How did the beta go in the U.S.?

    It went very well. We’ve done over 2,000 already. We shut it down. It’s enough to get feedback for the beta. I think we did 2,500 before we closed it and it’s not even a new phone.

    It’s half a year old, but it’s been much harder to get in the U.S.

    Yeah, but it doesn’t really support all the bands in the U.S., either. So you sometimes get 4G, or sometimes you get no signal.

    A limited release like that does serve the purpose of drumming up excitement. But given that the phone was already released in other parts of the world, what’s the purpose of the beta in the U.S.? Is it different from the feedback you get in Europe?

    We wanted more users to give feedback, and U.S., Europe and India users are different. I think the U.S. consumer is more focused on the experience, whereas Indian consumers more focus on the functionality, the feature set and the specs.

    \"\"

    Image Credits: Brian Heater

    What’s the difference between experience and functionality?

    U.S. users would be able to talk about user experience bugs, or what didn’t work, versus “Hey, why don’t we have this thing?”

    Why didn’t it launch in the U.S.?

    Engineering resources.

    For the bands?

    Okay, actually, two reasons. One is to add additional bands in the U.S. [but that] increases the cost of the product for all our regions around the world because we have a single SKU. So if we add everything we needed for the product, it became more expensive in Europe and India as well, but they had no benefit. And the second is our internal engineering. We just didn’t have enough engineering to support the different certifications that we needed to deal with some of the carriers.

    Are the bands similar in India and the U.K.?

    Yeah, basically.

    India is a huge market — the number two in the world. Is that why it was one of your first countries?

    There was a strong affinity for OnePlus in India. From a consumer side, we knew there was interest in what we were going to build, and from the partner side, from the sales channel side, they believe in our team based on what we did in the past, so they will support us again on this new journey.

    I see a good long-term potential, yeah. And then the market this year, India’s economy is going to grow by 6%. And a lot of economies are shrinking. They still have a lot of growth ahead of them as a nation. And also they have a very young population, and I think the middle class . . . will keep growing in size, whereas maybe in Europe, it will decrease in size.

    Apple and others focus their budget devices on the Indian market.  Did you consider doing a second variant?

    We did but I think that’s the easy way, but it also has long-term negative repercussions. For yourself as a company, if consumers only buy you because you’re cheap, then what’s your business model? A company needs to be profitable. So if you’re cheap today, then tomorrow somebody else can be even cheaper. And then this is like the fight for whoever makes the least amount of money or even loses money. So I think we’re taking a much more difficult route; we need to make our products different. And we need to create things that are useful for the customer. And we’ll know if it’s useful if they want to pay for it. It’ll be slower this way, but I think we’ll have a much stronger brand and much stronger product in the long-term if we do this, and also a healthier company, in terms of profitability.

    Do you have any operations in India?

    We have manufacturing in India. About half of the phones are made in India, and half in China. We’re building a manufacturing team in India to manage the factory. We have sales and marketing over there. And we have testing for some of our software and aftersales.

    The U.S. is looking to move more U.S. product manufacturing back to its own market. Is that something you would pursue at some point as a U.K. company, manufacturing in the U.K.?

    Probably not. I think this would need government policy support, and I don’t think the U.K. government has a plan there.

    The U.S. seems to be injecting a ton of money into that.

    I think India has a Made in India program, and the Middle East is starting to look at how they can diversify away from oil. I don’t think Europe in general has a plan for where they’re going.

    Pricing is obviously a big part of it. I see companies moving from Shenzhen to Vietnam or Mexico.

    I think the U.S. is not that expensive. Real estate for manufacturing is cheap compared to other places. The labor cost is higher, but the efficiency is also higher. So it kind of evens out.

    How aggressively are you looking at APAC?

    I think the China market is very competitive in an unhealthy way, as in it’s like a race to the bottom who can deliver the most amount of specs and features at the lowest possible price. And I think a majority are losing money in China.

    And you’d be competing with these huge companies with so many resources.

    I think, long-term, all the companies will become rational. When people are rational, we can consider entering the market. But it’s very irrational now. And also, if you want to operate in China, you need to have another software team because Google services are not there, so you need to build your own services. We just can’t think about all those things right now.

    \"\"

    Image Credits: Brian Heater

    What does your roadmap look like? OnePlus started small and has gradually grown its scope. Are you looking to ramp up the product release cadence for Nothing, too?

    Not very aggressively, because we want to create iconic products, and we want to support the products well, in terms of software. If you make like 10 or 20 phones a year, then it’s really hard to provide that level of support. It’s almost like you get a round of consumer interest for every new product. Like that always comes. So it’s kind of like a drug that you’re addicted to. You want to keep releasing stuff. But I think if you take the harder route, the more healthy route, they really figure out the product itself.

    Phone-wise, a yearly cadence makes sense.

    I can’t really comment on that right now. But I think if you look at the Phone (1), there’s actually still a sustained level of consumer interest, half a year after the launch. That’s becoming more and more rare in this industry.

    Any new product categories this year?

    Yes.

    \"Read



    >Source link>

    >>Join our Facebook page From top right corner. <<

  • Joe Biden ready to run for 2nd presidential term, U.S. first lady says: ‘He’s not done’ – National | Globalnews.ca

    U.S. First Lady Jill Biden has given one of the clearest indications yet that President Joe Biden will run for a second term. In an exclusive interview with The Associated Press, she said that there\’s \”pretty much\” nothing left to do but figure out the time and place for the announcement. Biden aides have said an announcement is likely to come in April. Jill Biden also spoke extensively for the first time about her skin cancer diagnosis, which led doctors to remove multiple basal cell lesions in January. She said she\’s been extra careful about sunscreen since then and has been advocating for cancer screening for years. The First Lady also reflected on the legacy of former President Jimmy Carter, who recently began home hospice care. She recalled him and his wife, Rosalynn, reaching out on the eve of Joe Biden’s inauguration two years ago and visiting the Carters at their home in Plains, Georgia, early in Biden’s presidency. As the First Lady continues her career in addition to her ceremonial duties, she expressed that she\’s not ready to think about retirement.



    Source link

    Join our Facebook page From top right corner.

  • Bidenworld chatters that Joe may not run

    [

    While the belief among nearly everyone in Biden’s orbit is that he’ll ultimately give the all-clear, the delay has resulted an awkward deep-freeze across the party — in which some potential presidential aspirants and scores of major donors are strategizing and even developing a Plan B while trying to remain respectful and publicly supportive of the 80-year-old president.

    Democratic Govs. JB Pritzker of Illinois, Gavin Newsom of California and Phil Murphy of New Jersey have taken steps that could be seen as aimed at keeping the door cracked if Biden bows out — though with enough ambiguity to give them plausible deniability. Senators like Bernie Sanders and Amy Klobuchar have been making similar moves.

    People directly in touch with the president described him as a kind of Hamlet on Delaware’s Christina River, warily biding his time as he ponders the particulars of his final campaign. In interviews, these people relayed an impression that the conventional wisdom in Washington, D.C. — that there’s simply no way he passes on 2024 — has crystallized too hard, too soon.

    “An inertia has set in,” one Biden confidant said. “It’s not that he won’t run, and the assumption is that he will. But nothing is decided. And it won’t be decided until it is.”

    ‘Doubts and problems if he waits’

    The stasis wasn’t always so pronounced. After former President Donald Trump’s launch in November, there was a desire among Biden advisers to begin charting their own kickoff plans in earnest. That urgency no longer is evident. They feel no threat of a credible primary challenge, a dynamic owed to Democrats’ better-than-expected midterms and a new early state presidential nominating calendar, handpicked by Biden. Holding off on signing campaign paperwork also allows Biden to avoid having to report a less-than-robust fundraising total for a first quarter that’s almost over.

    As the limbo continues, Biden’s advisers have been taking steps to staff a campaign and align with a top super PAC. Future Forward, which has been airing TV ads in support of the president’s agenda, would likely be Biden’s primary super PAC, though other groups would have a share in the campaign’s portfolio, a person familiar with the plans said.

    But to the surprise of some Biden allies, they say he has talked only sparingly about a possible campaign, three people familiar with the conversations said. His daily focus remains the job itself. Except for the occasional phone call with an adviser to review polling, he spends little time discussing the election. While First Lady Jill Biden signaled long ago she was on board with another run, some in the president’s orbit now wonder if the impending investigations into Hunter Biden could cause the president to second-guess a bid. Others believe it will not.

    A decision from Biden to forego another run would amount to a political earthquake not seen among Democrats in more than a half century, when Lyndon B. Johnson paired his partial halting of the U.S. bombing of Vietnam with his announcement to step aside, citing deepening “division in the American house now.”

    It would unleash an avalanche of attention on his vice president, Kamala Harris, whose uneven performances have raised doubts among fellow Democrats about her ability to win — either the primary, the general election, or both. And it would dislodge the logjam Biden himself created in 2020 when he dispatched with the sprawling field of Democratic contenders, a field that included Harris.

    “Obviously, it creates doubts and problems if he waits and waits and waits,” said Democratic strategist Mark Longabaugh, who continues to believe Biden will run — and that he won’t put off a decision for too long. “But if he were to somehow not declare ‘til June or something, I think some people would be stomping around.”

    “There would be a lot of negative conversation … among Democratic elites, and I just think that would force them to ultimately have to make a decision,” Longabaugh added. “I just don’t think he can dance around until sometime in the summer.”

    A campaign-in-waiting takes shape

    Biden and much of his inner circle still insists he plans to run, with the only caveat being a catastrophic health event that renders him unable. Anita Dunn, Jen O’Malley Dillon and Mike Donilon have effectively overseen the campaign-in-waiting, with Donilon considering shifting over to a campaign proper while the others manage operations from the White House.

    Other top advisers would also be heavily involved, including Steve Ricchetti and Bruce Reed, and former chief of staff Ron Klain may serve as an outside adviser for a 2024 bid.

    “The president has publicly told the country that he intends to run and has not made a final decision,” White House spokesman Andrew Bates said in a statement. “As you heard in the State of the Union, after the best midterm results for a new Democratic president in 60 years, his focus is on ‘finishing the job’ by delivering more results for American families and ensuring that our economy works from the bottom-up and the middle-out — not the top down.”

    For now, most of the senior team sees no need to rush, and are identifying April as the soonest he would go. That was the same month Biden unveiled his primary campaign in 2019, and the month that Barack Obama restarted his campaign engines in 2011. Bill Clinton declared in April of the year before he was reelected, and George W. Bush in May, Bates added.

    In addition to Biden’s unchallenged hold on the party, they note a belief that some of his legislative wins — like the infrastructure and CHIPS bills — will yield dividends in the months closer to Election Day and the need to pace the president. They point to the year ahead of heavy foreign travel, including his historic stops in Ukraine and Poland to rally European allies against Russia.

    “We’re not going to have a campaign until we have to,” a Biden adviser said. “He’s the president. Why does he need to dive into an election early?”

    But the delay in an announcement has allowed nervous chatter to seep in — or, in the case of Biden confidants, dribble out from his inner circle. It’s forced them to consider whether Biden’s waiting could leave the party in a difficult position should he opt against another run.

    Some people around the president note he’s always been, as he likes to say, somebody who respects fate. And they pointed to the seemingly unguarded answer he gave recently to Telemundo, when asked what was stopping him from announcing his decision on a second term.

    “I’m just not ready to make it,” Biden said. He continued to insist in the same interview that polls showing Democrats eager to move on from him are erroneous.

    Famously indecisive

    Biden is famously indecisive, a habit exacerbated by decades in the über-deliberative Senate. He publicly took his time mulling a decision not to run in 2016 and to launch his run in 2020. He missed two self-imposed deadlines before choosing Harris as a running-mate.

    In the White House, he pushed back the timeline to withdraw from Afghanistan; skipped over his initial benchmark to vaccinate 70 percent of American adults against Covid-19 with at least one shot; and earlier in his presidency let lapse deadlines on climate, commissions, mask standards and promised sanctions on Russia for poisoning opposition leader Alexei Navalny.

    His decision-making process is complete with extensive research, competing viewpoints and plenty of time to think. This time around, according to those close to him, he has made rounds of calls to longtime friends, all with an unspoken sense that he is running again — though without a firm commitment being made.

    Meanwhile, aspiring Democrats have moved to keep their options open. They’ve done so with enough ambiguity to give them cover — actions that could be interpreted as politicians simply running for reelection to a separate office, selling books, or building their profiles for a presidential campaign further out in the future.

    Among them is Pritzker, who was just elected to a second term. The Illinois Democrat — like everyone else — has offered his full support to Biden. But insiders note that senior advisers from his last two campaigns are still standing by just in case. Key among them is Quentin Fulks, who last year served as campaign manager to Georgia Sen. Raphael Warnock. Pritzker’s last two campaign managers, Mike Ollen, and chief of staff Anne Caprara, remain ready to deploy, along with others.

    “It’s the Boy Scout motto. ‘Be prepared,’” Democratic strategist David Axelrod said, referring to any appearance by Pritzker or other Democrats to be putting their ducks in a row for a potential presidential campaign.

    Newsom’s circle of top advisers and close aides have a similar understanding should he need to call on them — after easily winning reelection last year, surviving a recall attempt the year before and building one of the largest digital operations in Democratic politics. Murphy, who’s chairing the Democratic Governors Association, is in the same boat as the others, having vowed to back Biden while indicating an interest in a campaign should a lane open for him.

    Klobuchar (D-Minn.), who plans to seek reelection to her Senate seat in 2024, has been keeping up relations with donors far outside of Minnesota, holding a fundraiser in Philadelphia late last month. At the event, Klobuchar was asked if she planned on running for president in 2024, according to a person in the room. “She said she expects the president to run for reelection,” the person said.

    Sen. Elizabeth Warren (D-Mass.) also is running for reelection, a dynamic that allows her to pledge support for Biden, bank her own cash, communicate with party leaders on her own behalf — and change direction should she need to. One source close to the senator, however, said another presidential bid is highly unlikely regardless of what Biden decides.

    Sanders, who ran for the White House in 2020 and 2016, released a new book, “It’s OK to Be Angry About Capitalism,” this month. He is making media appearances and going on tour with stops in New York, Washington, D.C., Virginia, Arizona and California, the delegate-rich, Super Tuesday state that he won in his second presidential campaign.

    Sanders, who himself is 81, has said that he would not challenge Biden in a primary. But he had not ruled out a run in 2024 in the event there was an open presidential primary. Sanders’ former campaign co-chair, Rep. Ro Khanna (D-Calif), told POLITICO that Sanders “is preparing to run if Biden doesn’t,” adding he’d support Sanders in such a scenario.

    Khanna has made his own moves as well, retaining consultants in early-primary states and drawing contrasts with other ambitious Democrats such as former presidential candidate and Transportation Secretary Pete Buttigieg, another 2024 possibility. Khanna has said he will back Biden if he runs again and that he would not run for president next year if Biden declined to do so. But he has kept his options open to a campaign in 2028, or years beyond.

    “Without being overly aggressive, everyone’s still keeping the motor running just in case and they’re not being bashful about it,” said one Democratic donor, describing a call with the staff of a candidate who ran against Biden in 2020. “On the phone, everyone is very clear and has the same sentence up front: ‘If Joe Biden is running, no one will work harder than me, but if he’s not, for whatever reason, we just want to make sure we’re prepared for the good of the party.”

    The specter of Trump

    What’s driving the talk isn’t just Biden and his age, the donor added, but the possibility that Trump could return. “Most donors view the alternative as an existential threat to the country,” said the donor. “So is some of this impolite? Maybe. But no one seems to be taking issue with it.”

    As White House officials, advisers and operatives await word from Biden for 2024, many have received little clarity about where they may fit into an eventual campaign. Several decisions related to staffing remain up in the air — a dynamic some attribute to aides trying to best determine where all the moving pieces would fit together.

    Meanwhile, a plan to work in tandem with a constellation of Democratic super PACs is already starting to take shape.

    Dunn met in recent weeks with donors and officials at American Bridge, another major Democratic super PAC, one person familiar said. Top Biden aides have ties to both Future Forward and Priorities USA, two other super PACs.

    While Future Forward is likely to play the biggest role outside the possible campaign, aides stressed the others would be highly active, too. And it’s likely a campaign would designate an operative from outside its ranks itself to serve as an unofficial go-between to better coordinate with the outside groups.

    Several of the candidates for the campaign manager position represent a next generation of Democratic talent: Jennifer Ridder, Julie Chávez Rodriguez, Sam Cornale, Emma Brown and Preston Elliott. Christie Roberts, executive director of the Democratic Senate campaign arm and another sought-after operative, appears likely to remain in that job for 2024 following the party expanding its narrow Senate majority.

    Addisu Demissie, a longtime operative who ran Sen. Cory Booker’s 2020 campaign and worked closely with Bidenworld to produce the DNC, has been approached and courted for top posts on a campaign or super PAC. And Fulks, coming off the Warnock victory, also is viewed as a possible player on Biden’s campaign.

    Yet there are concerns about how much autonomy the role would provide given Biden’s tight-knit circle of old hands that’s famously suspicious of outsiders.

    There’s another complicating factor to sort out on staffing, according to the people familiar with the situation: Biden’s personal desire for a prominent campaign surrogate to blanket the cable airwaves.

    One person who could fit the bill of a more public-facing (less operationally involved) campaign manager is Kate Bedingfield, the Biden insider who just left her post as the White House communications director. Bedingfield’s name has come up more over the last week in conversations among Biden aides, the two people familiar with the talks said.

    The campaign pieces are being lined up. And several top financiers say they have been in touch with the president’s team to plan events. The president had a physical examination last week, in which his doctor gave him a nearly clean bill of health.

    All that is missing is the official go-ahead.

    Shia Kapos contributed to this report.



    Source link

    Join our Facebook page
    https://www.facebook.com/groups/www.pakistanaffairs.pk

  • OpenAI\’s Foundry will let customers buy dedicated compute to run its AI models

    OpenAI is quietly launching a new developer platform that lets customers run the company’s newer machine learning models, like GPT-3.5, on dedicated capacity. In screenshots of documentation published to Twitter by users with early access, OpenAI describes the forthcoming offering, called Foundry, as “designed for cutting-edge customers running larger workloads.”

    “[Foundry allows] inference at scale with full control over the model configuration and performance profile,” the documentation reads.

    If the screenshots are to be believed, Foundry — whenever it launches — will deliver a “static allocation” of compute capacity dedicated to a single customer. Users will be able to monitor specific instances with the same tools and dashboards that OpenAI uses to build and optimize models. In addition, Foundry will provide some level of version control, letting customers decide whether or not to upgrade to newer model releases, as well as “more robust” fine-tuning for OpenAI’s latest models.

    Foundry will also offer service-level commitments for instance uptime and on-calendar engineering support. Rentals will be based on dedicated compute units with three-month or one-year commitments; running an individual model instance will require a specific number of compute units (see the chart below).

    Instances won’t be cheap. Running a lightweight version of GPT-3.5 will cost $78,000 for a three-month commitment or $264,000 over a one-year commitment. To put that into perspective, one of Nvidia’s recent-gen supercomputers, the DGX Station, runs $149,000 per unit.

    Eagle-eyed Twitter and Reddit users spotted that one of the text-generating models listed in the instance pricing chart has a 32k max context window. (The context window refers to the text that the model considers before generating additional text; longer context windows allow the model to “remember” more text essentially.) GPT-3.5, OpenAI’s latest text-generating model, has a 4k max context window, suggesting that this mysterious new model could be the long-awaited GPT-4 — or a stepping stone toward it.

    OpenAI is under increasing pressure to turn a profit after a multi-billion-dollar investment from Microsoft. The company reportedly expects to make $200 million in 2023, a pittance compared to the more than $1 billion that’s been put toward the startup so far.

    Compute costs are largely to blame. Training state-of-the-art AI models can command upwards of millions of dollars, and running them generally isn’t much cheaper. According to OpenAI co-founder and CEO Sam Altman, it costs a few cents per chat to run ChatGPT, OpenAI’s viral chatbot — not an insignificant amount considering that ChatGPT had over a million users as of last December.

    In moves toward monetization, OpenAI recently launched a “pro” version of ChatGPT, ChatGPT Plus, starting at $20 per month and teamed up with Microsoft to develop Bing chat, a controversial chatbot (putting it mildly) that’s captured mainstream attention. According to Semafor and The Information, OpenAI plans to introduce a mobile ChatGPT app in the future and bring its AI language technology into Microsoft apps like Word, PowerPoint and Outlook.

    Separately, OpenAI continues to make its tech available through Microsoft’s Azure OpenAI Service, a business-focused model-serving platform, and maintain Copilot, a premium code-generating service developed in partnership with GitHub.





    Source link

    Join our Facebook page
    https://www.facebook.com/groups/www.pakistanaffairs.pk

  • OpenAI\’s Foundry will let customers buy dedicated compute to run its AI models

    OpenAI is quietly launching a new developer platform that lets customers run the company’s newer machine learning models, like GPT-3.5, on dedicated capacity. In screenshots of documentation published to Twitter by users with early access, OpenAI describes the forthcoming offering, called Foundry, as “designed for cutting-edge customers running larger workloads.”

    “[Foundry allows] inference at scale with full control over the model configuration and performance profile,” the documentation reads.

    If the screenshots are to be believed, Foundry — whenever it launches — will deliver a “static allocation” of compute capacity dedicated to a single customer. Users will be able to monitor specific instances with the same tools and dashboards that OpenAI uses to build and optimize models. In addition, Foundry will provide some level of version control, letting customers decide whether or not to upgrade to newer model releases, as well as “more robust” fine-tuning for OpenAI’s latest models.

    Foundry will also offer service-level commitments for instance uptime and on-calendar engineering support. Rentals will be based on dedicated compute units with three-month or one-year commitments; running an individual model instance will require a specific number of compute units (see the chart below).

    Instances won’t be cheap. Running a lightweight version of GPT-3.5 will cost $78,000 for a three-month commitment or $264,000 over a one-year commitment. To put that into perspective, one of Nvidia’s recent-gen supercomputers, the DGX Station, runs $149,000 per unit.

    Eagle-eyed Twitter and Reddit users spotted that one of the text-generating models listed in the instance pricing chart has a 32k max context window. (The context window refers to the text that the model considers before generating additional text; longer context windows allow the model to “remember” more text essentially.) GPT-3.5, OpenAI’s latest text-generating model, has a 4k max context window, suggesting that this mysterious new model could be the long-awaited GPT-4 — or a stepping stone toward it.

    OpenAI is under increasing pressure to turn a profit after a multi-billion-dollar investment from Microsoft. The company reportedly expects to make $200 million in 2023, a pittance compared to the more than $1 billion that’s been put toward the startup so far.

    Compute costs are largely to blame. Training state-of-the-art AI models can command upwards of millions of dollars, and running them generally isn’t much cheaper. According to OpenAI co-founder and CEO Sam Altman, it costs a few cents per chat to run ChatGPT, OpenAI’s viral chatbot — not an insignificant amount considering that ChatGPT had over a million users as of last December.

    In moves toward monetization, OpenAI recently launched a “pro” version of ChatGPT, ChatGPT Plus, starting at $20 per month and teamed up with Microsoft to develop Bing chat, a controversial chatbot (putting it mildly) that’s captured mainstream attention. According to Semafor and The Information, OpenAI plans to introduce a mobile ChatGPT app in the future and bring its AI language technology into Microsoft apps like Word, PowerPoint and Outlook.

    Separately, OpenAI continues to make its tech available through Microsoft’s Azure OpenAI Service, a business-focused model-serving platform, and maintain Copilot, a premium code-generating service developed in partnership with GitHub.





    Source link

    Join our Facebook page
    https://www.facebook.com/groups/www.pakistanaffairs.pk

  • OpenAI\’s Foundry will let customers buy dedicated compute to run its AI models

    OpenAI is quietly launching a new developer platform that lets customers run the company’s newer machine learning models, like GPT-3.5, on dedicated capacity. In screenshots of documentation published to Twitter by users with early access, OpenAI describes the forthcoming offering, called Foundry, as “designed for cutting-edge customers running larger workloads.”

    “[Foundry allows] inference at scale with full control over the model configuration and performance profile,” the documentation reads.

    If the screenshots are to be believed, Foundry — whenever it launches — will deliver a “static allocation” of compute capacity dedicated to a single customer. Users will be able to monitor specific instances with the same tools and dashboards that OpenAI uses to build and optimize models. In addition, Foundry will provide some level of version control, letting customers decide whether or not to upgrade to newer model releases, as well as “more robust” fine-tuning for OpenAI’s latest models.

    Foundry will also offer service-level commitments for instance uptime and on-calendar engineering support. Rentals will be based on dedicated compute units with three-month or one-year commitments; running an individual model instance will require a specific number of compute units (see the chart below).

    Instances won’t be cheap. Running a lightweight version of GPT-3.5 will cost $78,000 for a three-month commitment or $264,000 over a one-year commitment. To put that into perspective, one of Nvidia’s recent-gen supercomputers, the DGX Station, runs $149,000 per unit.

    Eagle-eyed Twitter and Reddit users spotted that one of the text-generating models listed in the instance pricing chart has a 32k max context window. (The context window refers to the text that the model considers before generating additional text; longer context windows allow the model to “remember” more text essentially.) GPT-3.5, OpenAI’s latest text-generating model, has a 4k max context window, suggesting that this mysterious new model could be the long-awaited GPT-4 — or a stepping stone toward it.

    OpenAI is under increasing pressure to turn a profit after a multi-billion-dollar investment from Microsoft. The company reportedly expects to make $200 million in 2023, a pittance compared to the more than $1 billion that’s been put toward the startup so far.

    Compute costs are largely to blame. Training state-of-the-art AI models can command upwards of millions of dollars, and running them generally isn’t much cheaper. According to OpenAI co-founder and CEO Sam Altman, it costs a few cents per chat to run ChatGPT, OpenAI’s viral chatbot — not an insignificant amount considering that ChatGPT had over a million users as of last December.

    In moves toward monetization, OpenAI recently launched a “pro” version of ChatGPT, ChatGPT Plus, starting at $20 per month and teamed up with Microsoft to develop Bing chat, a controversial chatbot (putting it mildly) that’s captured mainstream attention. According to Semafor and The Information, OpenAI plans to introduce a mobile ChatGPT app in the future and bring its AI language technology into Microsoft apps like Word, PowerPoint and Outlook.

    Separately, OpenAI continues to make its tech available through Microsoft’s Azure OpenAI Service, a business-focused model-serving platform, and maintain Copilot, a premium code-generating service developed in partnership with GitHub.





    Source link

    Join our Facebook page
    https://www.facebook.com/groups/www.pakistanaffairs.pk

  • OpenAI\’s Foundry will let customers buy dedicated compute to run its AI models

    OpenAI is quietly launching a new developer platform that lets customers run the company’s newer machine learning models, like GPT-3.5, on dedicated capacity. In screenshots of documentation published to Twitter by users with early access, OpenAI describes the forthcoming offering, called Foundry, as “designed for cutting-edge customers running larger workloads.”

    “[Foundry allows] inference at scale with full control over the model configuration and performance profile,” the documentation reads.

    If the screenshots are to be believed, Foundry — whenever it launches — will deliver a “static allocation” of compute capacity dedicated to a single customer. Users will be able to monitor specific instances with the same tools and dashboards that OpenAI uses to build and optimize models. In addition, Foundry will provide some level of version control, letting customers decide whether or not to upgrade to newer model releases, as well as “more robust” fine-tuning for OpenAI’s latest models.

    Foundry will also offer service-level commitments for instance uptime and on-calendar engineering support. Rentals will be based on dedicated compute units with three-month or one-year commitments; running an individual model instance will require a specific number of compute units (see the chart below).

    Instances won’t be cheap. Running a lightweight version of GPT-3.5 will cost $78,000 for a three-month commitment or $264,000 over a one-year commitment. To put that into perspective, one of Nvidia’s recent-gen supercomputers, the DGX Station, runs $149,000 per unit.

    Eagle-eyed Twitter and Reddit users spotted that one of the text-generating models listed in the instance pricing chart has a 32k max context window. (The context window refers to the text that the model considers before generating additional text; longer context windows allow the model to “remember” more text essentially.) GPT-3.5, OpenAI’s latest text-generating model, has a 4k max context window, suggesting that this mysterious new model could be the long-awaited GPT-4 — or a stepping stone toward it.

    OpenAI is under increasing pressure to turn a profit after a multi-billion-dollar investment from Microsoft. The company reportedly expects to make $200 million in 2023, a pittance compared to the more than $1 billion that’s been put toward the startup so far.

    Compute costs are largely to blame. Training state-of-the-art AI models can command upwards of millions of dollars, and running them generally isn’t much cheaper. According to OpenAI co-founder and CEO Sam Altman, it costs a few cents per chat to run ChatGPT, OpenAI’s viral chatbot — not an insignificant amount considering that ChatGPT had over a million users as of last December.

    In moves toward monetization, OpenAI recently launched a “pro” version of ChatGPT, ChatGPT Plus, starting at $20 per month and teamed up with Microsoft to develop Bing chat, a controversial chatbot (putting it mildly) that’s captured mainstream attention. According to Semafor and The Information, OpenAI plans to introduce a mobile ChatGPT app in the future and bring its AI language technology into Microsoft apps like Word, PowerPoint and Outlook.

    Separately, OpenAI continues to make its tech available through Microsoft’s Azure OpenAI Service, a business-focused model-serving platform, and maintain Copilot, a premium code-generating service developed in partnership with GitHub.





    Source link

    Join our Facebook page
    https://www.facebook.com/groups/www.pakistanaffairs.pk